Building & Evaluating Predictive model: Supermarket Business CaseSiddhanth Chaurasiya
- The document describes building predictive models using decision tree and regression modeling to predict which customers are likely to purchase new organic products being introduced by a supermarket.
- Both decision tree and logistic regression models were created, with the decision tree models performing slightly better based on various evaluation metrics such as cumulative lift, ROC curve, and average square error.
- The top variables influencing the likelihood of a customer purchasing organics according to the models were gender, age, and affluence level.
Bank - Loan Purchase Modeling
This case is about a bank which has a growing customer base. Majority of these customers are liability customers (depositors) with varying size of deposits. The number of customers who are also borrowers (asset customers) is quite small, and the bank is interested in expanding this base rapidly to bring in more loan business and in the process, earn more through the interest on loans. In particular, the management wants to explore ways of converting its liability customers to personal loan customers (while retaining them as depositors). A campaign that the bank ran last year for liability customers showed a healthy conversion rate of over 9% success. This has encouraged the retail marketing department to devise campaigns with better target marketing to increase the success ratio with a minimal budget. The department wants to build a model that will help them identify the potential customers who have a higher probability of purchasing the loan. This will increase the success ratio while at the same time reduce the cost of the campaign. The dataset has data on 5000 customers. The data include customer demographic information (age, income, etc.), the customer's relationship with the bank (mortgage, securities account, etc.), and the customer response to the last personal loan campaign (Personal Loan). Among these 5000 customers, only 480 (= 9.6%) accepted the personal loan that was offered to them in the earlier campaign.
Our job is to build the best model which can classify the right customers who have a higher probability of purchasing the loan. We are expected to do the following:
EDA of the data available. Showcase the results using appropriate graphs
Apply appropriate clustering on the data and interpret the output .
Build appropriate models on both the test and train data (CART & Random Forest). Interpret all the model outputs and do the necessary modifications wherever eligible (such as pruning).
Check the performance of all the models that you have built (test and train). Use all the model performance measures you have learned so far. Share your remarks on which model performs the best.
The document summarizes the HeC CaseReasoner system, which uses neighborhood graphs to represent clinical case data and patient similarity for clinical decision support. It describes how discriminative distance functions can be learned from equivalence constraints or using intrinsic random forest distances to combine the transparency of case retrieval with the power of machine learning models. The system provides an interactive graphical interface for clinicians to explore patient records and visualizations to aid clinical decision making.
Through this project, we observed that a larger dataset with higher skewness in data distribution trained the model better and faster than a smaller dataset with lesser skewness in data distribution.
Ben-Graham Lighting pre-processing technique resulted in better model performance. The model was also able to learn the features of interest such as haemorrhages, hard exudates etc.
ResNet-50 and ResNet-101 models gave the best performance out of all the models we experimented with.
The first few initial layers of both ResNet-50 and ResNet-101 yielded high accuracy of predictions, especially that of ResNet-101 which was almost equal to the final accuracy obtained at the end of all layers.
The model was robust to almost all types of masking for grade ‘0’ images vowing to the higher number of data points in the training dataset, whereas other classes of images were somewhat sensitive to the maskings near the middle portion of the fundus image.
Images centred around the fovea, optic disc and centre contributed significantly to the classification, among which those centred around the fovea produced very close performance similar to that of full fundus images.
A NOVEL DENSITY-BASED CLUSTERING ALGORITHM FOR PREDICTING CARDIOVASCULAR DISEASEindexPub
Cardiovascular diseases (CVDs) remain a leading cause of global morbidity and mortality. Early identification of individuals at risk of heart disease is crucial for effective preventive interventions. To improve the prediction accuracy, this paper proposed Heart Disease Prediction using the Density-Based Ordering of Clustering Objects (DBOCO) framework. The Dataset has been pre-processed using Weighted Transform K-Means Clustering (WTKMC). Features are selected using Ensemble Feature Selection (EFS) with a Weighted Binary Bat Algorithm (WBBAT) used to ensure that the emphasis is on the most relevant predictors. Finally, the prediction has been done using the Density-Based Ordering of Clustering method, which has been designed exclusively for cardiovascular disease prediction. DBOCO, a density-based clustering approach, effectively finds dense clusters within data, allowing for the inherent overlap in cardiovascular risk variables. DBOCO captures complicated patterns by detecting these overlapping clusters, improving the accuracy of disease prediction models. The proposed approach has been verified with heart disease datasets, displaying higher performance than traditional methods. This study marks a substantial leap in predicting cardiovascular disease providing a comprehensive and dependable framework for early identification and preventive concern.
Algoritma Random Forest beserta aplikasi nyabatubao
Random forest is an ensemble classifier that consists of many decision trees. It outputs the class that is the mode of the classes from individual trees. Each tree is constructed by selecting a random sample of training cases and a small random subset of input variables. Trees are fully grown and not pruned, and each tree votes for the most popular class. The random forest algorithm averages these votes for classification or averages predictions for regression. Random forests have advantages such as high accuracy, efficiency with large datasets, and estimates of variable importance.
Building & Evaluating Predictive model: Supermarket Business CaseSiddhanth Chaurasiya
- The document describes building predictive models using decision tree and regression modeling to predict which customers are likely to purchase new organic products being introduced by a supermarket.
- Both decision tree and logistic regression models were created, with the decision tree models performing slightly better based on various evaluation metrics such as cumulative lift, ROC curve, and average square error.
- The top variables influencing the likelihood of a customer purchasing organics according to the models were gender, age, and affluence level.
Bank - Loan Purchase Modeling
This case is about a bank which has a growing customer base. Majority of these customers are liability customers (depositors) with varying size of deposits. The number of customers who are also borrowers (asset customers) is quite small, and the bank is interested in expanding this base rapidly to bring in more loan business and in the process, earn more through the interest on loans. In particular, the management wants to explore ways of converting its liability customers to personal loan customers (while retaining them as depositors). A campaign that the bank ran last year for liability customers showed a healthy conversion rate of over 9% success. This has encouraged the retail marketing department to devise campaigns with better target marketing to increase the success ratio with a minimal budget. The department wants to build a model that will help them identify the potential customers who have a higher probability of purchasing the loan. This will increase the success ratio while at the same time reduce the cost of the campaign. The dataset has data on 5000 customers. The data include customer demographic information (age, income, etc.), the customer's relationship with the bank (mortgage, securities account, etc.), and the customer response to the last personal loan campaign (Personal Loan). Among these 5000 customers, only 480 (= 9.6%) accepted the personal loan that was offered to them in the earlier campaign.
Our job is to build the best model which can classify the right customers who have a higher probability of purchasing the loan. We are expected to do the following:
EDA of the data available. Showcase the results using appropriate graphs
Apply appropriate clustering on the data and interpret the output .
Build appropriate models on both the test and train data (CART & Random Forest). Interpret all the model outputs and do the necessary modifications wherever eligible (such as pruning).
Check the performance of all the models that you have built (test and train). Use all the model performance measures you have learned so far. Share your remarks on which model performs the best.
The document summarizes the HeC CaseReasoner system, which uses neighborhood graphs to represent clinical case data and patient similarity for clinical decision support. It describes how discriminative distance functions can be learned from equivalence constraints or using intrinsic random forest distances to combine the transparency of case retrieval with the power of machine learning models. The system provides an interactive graphical interface for clinicians to explore patient records and visualizations to aid clinical decision making.
Through this project, we observed that a larger dataset with higher skewness in data distribution trained the model better and faster than a smaller dataset with lesser skewness in data distribution.
Ben-Graham Lighting pre-processing technique resulted in better model performance. The model was also able to learn the features of interest such as haemorrhages, hard exudates etc.
ResNet-50 and ResNet-101 models gave the best performance out of all the models we experimented with.
The first few initial layers of both ResNet-50 and ResNet-101 yielded high accuracy of predictions, especially that of ResNet-101 which was almost equal to the final accuracy obtained at the end of all layers.
The model was robust to almost all types of masking for grade ‘0’ images vowing to the higher number of data points in the training dataset, whereas other classes of images were somewhat sensitive to the maskings near the middle portion of the fundus image.
Images centred around the fovea, optic disc and centre contributed significantly to the classification, among which those centred around the fovea produced very close performance similar to that of full fundus images.
A NOVEL DENSITY-BASED CLUSTERING ALGORITHM FOR PREDICTING CARDIOVASCULAR DISEASEindexPub
Cardiovascular diseases (CVDs) remain a leading cause of global morbidity and mortality. Early identification of individuals at risk of heart disease is crucial for effective preventive interventions. To improve the prediction accuracy, this paper proposed Heart Disease Prediction using the Density-Based Ordering of Clustering Objects (DBOCO) framework. The Dataset has been pre-processed using Weighted Transform K-Means Clustering (WTKMC). Features are selected using Ensemble Feature Selection (EFS) with a Weighted Binary Bat Algorithm (WBBAT) used to ensure that the emphasis is on the most relevant predictors. Finally, the prediction has been done using the Density-Based Ordering of Clustering method, which has been designed exclusively for cardiovascular disease prediction. DBOCO, a density-based clustering approach, effectively finds dense clusters within data, allowing for the inherent overlap in cardiovascular risk variables. DBOCO captures complicated patterns by detecting these overlapping clusters, improving the accuracy of disease prediction models. The proposed approach has been verified with heart disease datasets, displaying higher performance than traditional methods. This study marks a substantial leap in predicting cardiovascular disease providing a comprehensive and dependable framework for early identification and preventive concern.
Algoritma Random Forest beserta aplikasi nyabatubao
Random forest is an ensemble classifier that consists of many decision trees. It outputs the class that is the mode of the classes from individual trees. Each tree is constructed by selecting a random sample of training cases and a small random subset of input variables. Trees are fully grown and not pruned, and each tree votes for the most popular class. The random forest algorithm averages these votes for classification or averages predictions for regression. Random forests have advantages such as high accuracy, efficiency with large datasets, and estimates of variable importance.
This document discusses a project that uses machine learning algorithms to predict potential heart diseases. The project uses a dataset with 13 features and applies algorithms like K-Nearest Neighbors Classifier and Support Vector Classifier, with and without PCA. The K-Nearest Neighbors Classifier achieved the best accuracy score of 87% at predicting heart disease based on the dataset.
This document provides information about getting fully solved assignments for various postgraduate programs and semesters. Students can send their semester and specialization details to the provided email ID or call the given phone number to get assignments. It includes details of subject codes, credits, and marks for assignments related to research methodology for programs like MBA, PGDM, PGDHRM etc. for semesters 1 and 3.
Clustering Medical Data to Predict the Likelihood of Diseasesrazanpaul
This document proposes a constraint k-Means-Mode clustering algorithm to predict the likelihood of diseases using medical data containing both continuous and categorical attributes. It first maps complex medical data to mineable items using domain dictionaries and rule bases. The developed algorithm can handle both continuous and discrete data, perform clustering based on anticipated likelihood attributes with core disease attributes, and was tested on a real-world patient dataset to demonstrate its effectiveness.
This document discusses a method for splitting large medical data sets based on the normal distribution in a cloud computing environment. The key points are:
- Large medical and e-commerce data sets present challenges for data mining due to their size and generation velocity. Existing splitting methods like UV decomposition do not scale well for very large data sets.
- The proposed method splits large data sets into smaller subsets based on identifying groups of data that approximate a normal distribution. These normal distribution (ND) subsets can then be analyzed individually while still representing the overall data set.
- The ND subsets are well-suited for distributed processing in a cloud computing environment, as each subset can be analyzed locally and in parallel. Experimental results show
Cluster analysis is an unsupervised machine learning technique that groups unlabeled data points into clusters based on similarities. It aims to divide data into meaningful groups (clusters) where items in the same cluster are more similar to each other than items in different clusters. The document discusses different types of cluster analysis methods including hierarchical agglomerative clustering which starts with each data point as its own cluster and merges them together based on similarity, and k-means clustering which partitions data into k mutually exclusive clusters where each observation belongs to the cluster with the nearest mean. It also covers topics like distance measures, selecting the optimal number of clusters, and limitations of cluster analysis techniques.
This document provides a tutorial on conducting and interpreting a multiple linear regression analysis in SPSS. It contains two sections - the first outlines the steps to specify a regression analysis in SPSS using sample data. The second section interprets example SPSS output, including descriptive statistics, bivariate correlations, model summary, ANOVA table, and coefficients output. It also provides a guide for writing up the results in APA style.
This document summarizes a heart disease data analysis project. It discusses data analysis steps like data exploration, cleaning, and model building. The analysis uses a heart disease dataset from Kaggle with 13 independent variables and 1 dependent variable indicating the presence or absence of heart disease. Various algorithms are tested on training and test splits of the data, with random forest classification found to have the best accuracy in predicting heart disease.
The document discusses principal component analysis (PCA) and linear discriminant analysis (LDA) for dimensionality reduction in pattern recognition and their application to face recognition. PCA finds the directions along which the data varies the most to reduce dimensionality while retaining variation. LDA seeks directions that maximize between-class variation and minimize within-class variation. Studies show LDA performs better than PCA for classification when the training set is large and representative of each class.
CLUSTERING DICHOTOMOUS DATA FOR HEALTH CAREijistjournal
This document discusses clustering dichotomous health care data using the K-means algorithm after transforming the data using Wiener transformation. It begins with an introduction to dichotomous data and the challenges of clustering medical data. It then describes the K-means clustering algorithm and various distance measures used for binary data clustering. The document proposes using Wiener transformation to first transform binary data to real values before applying K-means clustering. It evaluates the results on a lens dataset using inter-cluster and intra-cluster distances, finding the transformed data yields better clusters than the original binary data according to these metrics.
The document presents a study that uses machine learning techniques to build a diagnostic model to distinguish between very mild dementia (VMD) and cognitively normal individuals using MRI data. Seven machine learning algorithms were tested including naive Bayes, Bayesian networks, decision trees, support vector machines, and neural networks. The right hippocampus was the most important discriminating brain region. Algorithms like naive Bayes and support vector machines performed better than previous statistical approaches at classifying VMD versus controls based on MRI data. Cross-validation is a more reliable performance measure than accuracy alone.
Comparison on PCA ICA and LDA in Face Recognitionijdmtaiir
Face recognition is used in wide range of application.
In recent years, face recognition has become one of the most
successful applications in image analysis and understanding.
Different statistical method and research groups reported a
contradictory result when comparing principal component
analysis (PCA) algorithm, independent component analysis
(ICA) algorithm, and linear discriminant analysis (LDA)
algorithm that has been proposed in recent years. The goal of
this paper is to compare and analyze the three algorithms and
conclude which is best. Feret Dataset is used for consistency
Data Samples & Data AnalysesNYU SCPSDatabaOllieShoresna
Data Samples & Data Analyses
NYU | SCPS
Database Management & Modeling
Edward Colet
[email protected]
Asynchronous Session 3, week of June 7 2021
Class material and homework so far
You should be through text chapters 1-3 (introduction), 4-5 (database fundamentals), and the supplemental readings on RDMBS’s and BigData;
HW submissions were a short write-up about yourselves (hw1), a relational database design exercise, and a BigData discussion (hw2)
Questions from the material?
Please feel free to also use the discussion section on our NYU Discussion site to ask, answer, comment on material from this week (for this week, this will be part of hw3)
Content for this week
Chapter 6: The Analysis Sample
Chapter 7: Analyzing and Manipulating Customer Data
Online Khan Academy content to Introduce SQL
Week3 Overview
‹#›
‹#›
Key themes for this week (and the course)
Databases are important for storing data (obviously), but you have to analyze the data as well otherwise you just have a “data tomb”. The analysis of data to gain insights is what gives the data it’s power and makes it really valuable.
This week we’ll learn about some fundamental analytic concepts operations associated with analysis; We’ll review Correlation, a foundational basis for analytics and modeling; We’ll learn some of the fundamental operations to slice and dice data, and we’ll write basic SQL (Structured Query Language) code to create a table, populate it with records, and query the table to extract and summarize information.
Week3 Overview
‹#›
‹#›
The Analysis Sample
Chapter 6
Key Point of the Chapter:
Data analyses are usually (almost always) done on subsets of the data in the database. As such, the following are the key concepts and points to understand about working with subsets of data
Representative samples
Random samples
Frozen files
Test and validation data sets
Chapter 6: The Analysis Sample
‹#›
‹#›
Know some common marketing scenarios that would be suitable to use a sample and test…
To gauge new product offering/response
Price elasticity
Impact of a creative / change
Identify target market for new test
Gain insights on specific groups/segments
. . . Any decision about your product in the market can be tested and analyzed to minimize/gauge risk
Chapter 6: The Analysis Sample
‹#›
‹#›
What is a representative sample?
A sample accurately reflecting the population of interest from which the marketer wants to draw inferences.
Can not extend or apply results from one population to another
Can not purposely exclude names except for “permission opt-outs”, or other recently promoted per rules/regulations
What is a random sample?
When every member equally likely to be chosen
Nth selects is one approach (select every nth name)
Chapter 6: The Analysis Sample
‹#›
‹#›
What is a frozen file?
A file containing a snapshot view of the customer(s) at the time of the promotion, updated with response data to the promotion
Why is a fr ...
[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...Ziyuan Zhao
Slides presented at International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2022, Singapore. DOI: 10.1007/978-3-031-16443-9_13
This presentation discusses analyzing and predicting heart disease using machine learning techniques. The main goal is to build a model to predict heart disease occurrence based on risk factors. Several classification algorithms will be tested, including K-nearest neighbors, decision trees, logistic regression, naive Bayes, support vector machines, and random forests. Publicly available datasets with over 400 cases will be used to train and evaluate the models. Existing literature has used techniques like association rules, neural networks, and clustering for heart disease prediction, achieving accuracy between 70-90%. The proposed approach aims to improve prediction performance.
This document provides an introduction to multivariate data analysis (MVA) using R. It defines what multivariate analysis is and explains that multivariate datasets contain multiple variables and can include mixed data types. Common MVA methods are discussed, including hierarchical cluster analysis (HCA) and partition clustering for exploratory analysis. HCA involves calculating distances between variables, building a tree to visualize relationships, and can identify potential subgroups. Partition clustering assigns variables to discrete clusters. The document demonstrates HCA and partition clustering on gene expression data to explore patterns among patients and genes.
CLUSTERING DICHOTOMOUS DATA FOR HEALTH CAREijistjournal
Dichotomous data is a type of categorical data, which is binary with categories zero and one. Health care data is one of the heavily used categorical data. Binary data are the simplest form of data used for heath care databases in which close ended questions can be used; it is very efficient based on computational efficiency and memory capacity to represent categorical type data. Clustering health care or medical data is very tedious due to its complex data representation models, high dimensionality and data sparsity. In this paper, clustering is performed after transforming the dichotomous data into real by wiener transformation. The proposed algorithm can be usable for determining the correlation of the health disorders and symptoms observed in large medical and health binary databases. Computational results show that the clustering based on Wiener transformation is very efficient in terms of objectivity and subjectivity.
The document discusses data analytics solutions involving machine learning and statistical modeling. It proposes splitting the solution into two parts: 1) applying algorithm techniques and statistical tests to data, and 2) making data-driven decisions using insights, metrics, and innovations. It then provides more details on machine learning techniques like training/testing data sets, and determining the optimal number of neurons in neural networks. Statistical modeling techniques like logistic regression, decision trees, and neural networks are recommended. The document emphasizes comparing different model results to identify ways to improve performance.
1) The document discusses using various machine learning and predictive analytics techniques like random forests, support vector machines, neural networks, and network analysis to analyze healthcare claims data and detect anomalies and fraud.
2) Over 87 million claim lines from 17,000 providers and 1.6 million members are analyzed monthly using 70 measures of claiming behavior and 6 algorithms. Small clusters of providers with high-risk claiming patterns are identified.
3) Random forest classification and support vector machines are used to predict which investigations will yield findings, and random forests provide a measure of variable importance. Network analysis can reveal problematic relationships between providers and members.
Comparative Analysis of Weighted Emphirical Optimization Algorithm and Lazy C...IIRindia
Health care has millions of centric data to discover the essential data is more important. In data mining the discovery of hidden information can be more innovative and useful for much necessity constraint in the field of forecasting, patient’s behavior, executive information system, e-governance the data mining tools and technique play a vital role. In Parkinson health care domain the hidden concept predicts the possibility of likelihood of the disease and also ensures the important feature attribute. The explicit patterns are converted to implicit by applying various algorithms i.e., association, clustering, classification to arrive at the full potential of the medical data. In this research work Parkinson dataset have been used with different classifiers to estimate the accuracy, sensitivity, specificity, kappa and roc characteristics. The proposed weighted empirical optimization algorithm is compared with other classifiers to be efficient in terms of accuracy and other related measures. The proposed model exhibited utmost accuracy of 87.17% with a robust kappa statistics measurement and roc degree indicated the strong stability of the model when compared to other classifiers. The total penalty cost generated by the proposed model is less when compared with the penalty cost of other classifiers in addition to accuracy and other performance measures.
Lead dbs Workshop 2020 Brisbane ProgrammeAndreas Horn
The document describes a 2-day workshop on deep brain stimulation (DBS) neuroimaging techniques using the Lead-DBS software package. The workshop will cover topics like electrode localization, spatial normalization, connectivity mapping, and will provide hands-on sessions for participants to practice techniques on their own data. Attendees should have some experience with MATLAB and Lead-DBS and bring their own laptops with required software installed. The agenda includes sessions on imaging pipelines, connectomics, troubleshooting, and group analysis methods.
- Deep brain stimulation (DBS) of the subthalamic nucleus (STN) in Parkinson's disease patients affects both motor and non-motor functions through interactions with motor, associative, and limbic networks in the basal ganglia.
- Connectomics analysis using fiber tracking from DBS electrode locations can help explain individual variability in behavioral effects of STN-DBS across different cognitive tasks. Specifically, it shows that stimulation of fibers connecting regions like the pre-SMA can predict detriments in stopping behavior.
- Stimulation of prefrontal fibers bypassing the STN that connect to brainstem regions has been linked to worsening of depressive symptoms after surgery through connectomics analysis.
This document discusses a project that uses machine learning algorithms to predict potential heart diseases. The project uses a dataset with 13 features and applies algorithms like K-Nearest Neighbors Classifier and Support Vector Classifier, with and without PCA. The K-Nearest Neighbors Classifier achieved the best accuracy score of 87% at predicting heart disease based on the dataset.
This document provides information about getting fully solved assignments for various postgraduate programs and semesters. Students can send their semester and specialization details to the provided email ID or call the given phone number to get assignments. It includes details of subject codes, credits, and marks for assignments related to research methodology for programs like MBA, PGDM, PGDHRM etc. for semesters 1 and 3.
Clustering Medical Data to Predict the Likelihood of Diseasesrazanpaul
This document proposes a constraint k-Means-Mode clustering algorithm to predict the likelihood of diseases using medical data containing both continuous and categorical attributes. It first maps complex medical data to mineable items using domain dictionaries and rule bases. The developed algorithm can handle both continuous and discrete data, perform clustering based on anticipated likelihood attributes with core disease attributes, and was tested on a real-world patient dataset to demonstrate its effectiveness.
This document discusses a method for splitting large medical data sets based on the normal distribution in a cloud computing environment. The key points are:
- Large medical and e-commerce data sets present challenges for data mining due to their size and generation velocity. Existing splitting methods like UV decomposition do not scale well for very large data sets.
- The proposed method splits large data sets into smaller subsets based on identifying groups of data that approximate a normal distribution. These normal distribution (ND) subsets can then be analyzed individually while still representing the overall data set.
- The ND subsets are well-suited for distributed processing in a cloud computing environment, as each subset can be analyzed locally and in parallel. Experimental results show
Cluster analysis is an unsupervised machine learning technique that groups unlabeled data points into clusters based on similarities. It aims to divide data into meaningful groups (clusters) where items in the same cluster are more similar to each other than items in different clusters. The document discusses different types of cluster analysis methods including hierarchical agglomerative clustering which starts with each data point as its own cluster and merges them together based on similarity, and k-means clustering which partitions data into k mutually exclusive clusters where each observation belongs to the cluster with the nearest mean. It also covers topics like distance measures, selecting the optimal number of clusters, and limitations of cluster analysis techniques.
This document provides a tutorial on conducting and interpreting a multiple linear regression analysis in SPSS. It contains two sections - the first outlines the steps to specify a regression analysis in SPSS using sample data. The second section interprets example SPSS output, including descriptive statistics, bivariate correlations, model summary, ANOVA table, and coefficients output. It also provides a guide for writing up the results in APA style.
This document summarizes a heart disease data analysis project. It discusses data analysis steps like data exploration, cleaning, and model building. The analysis uses a heart disease dataset from Kaggle with 13 independent variables and 1 dependent variable indicating the presence or absence of heart disease. Various algorithms are tested on training and test splits of the data, with random forest classification found to have the best accuracy in predicting heart disease.
The document discusses principal component analysis (PCA) and linear discriminant analysis (LDA) for dimensionality reduction in pattern recognition and their application to face recognition. PCA finds the directions along which the data varies the most to reduce dimensionality while retaining variation. LDA seeks directions that maximize between-class variation and minimize within-class variation. Studies show LDA performs better than PCA for classification when the training set is large and representative of each class.
CLUSTERING DICHOTOMOUS DATA FOR HEALTH CAREijistjournal
This document discusses clustering dichotomous health care data using the K-means algorithm after transforming the data using Wiener transformation. It begins with an introduction to dichotomous data and the challenges of clustering medical data. It then describes the K-means clustering algorithm and various distance measures used for binary data clustering. The document proposes using Wiener transformation to first transform binary data to real values before applying K-means clustering. It evaluates the results on a lens dataset using inter-cluster and intra-cluster distances, finding the transformed data yields better clusters than the original binary data according to these metrics.
The document presents a study that uses machine learning techniques to build a diagnostic model to distinguish between very mild dementia (VMD) and cognitively normal individuals using MRI data. Seven machine learning algorithms were tested including naive Bayes, Bayesian networks, decision trees, support vector machines, and neural networks. The right hippocampus was the most important discriminating brain region. Algorithms like naive Bayes and support vector machines performed better than previous statistical approaches at classifying VMD versus controls based on MRI data. Cross-validation is a more reliable performance measure than accuracy alone.
Comparison on PCA ICA and LDA in Face Recognitionijdmtaiir
Face recognition is used in wide range of application.
In recent years, face recognition has become one of the most
successful applications in image analysis and understanding.
Different statistical method and research groups reported a
contradictory result when comparing principal component
analysis (PCA) algorithm, independent component analysis
(ICA) algorithm, and linear discriminant analysis (LDA)
algorithm that has been proposed in recent years. The goal of
this paper is to compare and analyze the three algorithms and
conclude which is best. Feret Dataset is used for consistency
Data Samples & Data AnalysesNYU SCPSDatabaOllieShoresna
Data Samples & Data Analyses
NYU | SCPS
Database Management & Modeling
Edward Colet
[email protected]
Asynchronous Session 3, week of June 7 2021
Class material and homework so far
You should be through text chapters 1-3 (introduction), 4-5 (database fundamentals), and the supplemental readings on RDMBS’s and BigData;
HW submissions were a short write-up about yourselves (hw1), a relational database design exercise, and a BigData discussion (hw2)
Questions from the material?
Please feel free to also use the discussion section on our NYU Discussion site to ask, answer, comment on material from this week (for this week, this will be part of hw3)
Content for this week
Chapter 6: The Analysis Sample
Chapter 7: Analyzing and Manipulating Customer Data
Online Khan Academy content to Introduce SQL
Week3 Overview
‹#›
‹#›
Key themes for this week (and the course)
Databases are important for storing data (obviously), but you have to analyze the data as well otherwise you just have a “data tomb”. The analysis of data to gain insights is what gives the data it’s power and makes it really valuable.
This week we’ll learn about some fundamental analytic concepts operations associated with analysis; We’ll review Correlation, a foundational basis for analytics and modeling; We’ll learn some of the fundamental operations to slice and dice data, and we’ll write basic SQL (Structured Query Language) code to create a table, populate it with records, and query the table to extract and summarize information.
Week3 Overview
‹#›
‹#›
The Analysis Sample
Chapter 6
Key Point of the Chapter:
Data analyses are usually (almost always) done on subsets of the data in the database. As such, the following are the key concepts and points to understand about working with subsets of data
Representative samples
Random samples
Frozen files
Test and validation data sets
Chapter 6: The Analysis Sample
‹#›
‹#›
Know some common marketing scenarios that would be suitable to use a sample and test…
To gauge new product offering/response
Price elasticity
Impact of a creative / change
Identify target market for new test
Gain insights on specific groups/segments
. . . Any decision about your product in the market can be tested and analyzed to minimize/gauge risk
Chapter 6: The Analysis Sample
‹#›
‹#›
What is a representative sample?
A sample accurately reflecting the population of interest from which the marketer wants to draw inferences.
Can not extend or apply results from one population to another
Can not purposely exclude names except for “permission opt-outs”, or other recently promoted per rules/regulations
What is a random sample?
When every member equally likely to be chosen
Nth selects is one approach (select every nth name)
Chapter 6: The Analysis Sample
‹#›
‹#›
What is a frozen file?
A file containing a snapshot view of the customer(s) at the time of the promotion, updated with response data to the promotion
Why is a fr ...
[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...Ziyuan Zhao
Slides presented at International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2022, Singapore. DOI: 10.1007/978-3-031-16443-9_13
This presentation discusses analyzing and predicting heart disease using machine learning techniques. The main goal is to build a model to predict heart disease occurrence based on risk factors. Several classification algorithms will be tested, including K-nearest neighbors, decision trees, logistic regression, naive Bayes, support vector machines, and random forests. Publicly available datasets with over 400 cases will be used to train and evaluate the models. Existing literature has used techniques like association rules, neural networks, and clustering for heart disease prediction, achieving accuracy between 70-90%. The proposed approach aims to improve prediction performance.
This document provides an introduction to multivariate data analysis (MVA) using R. It defines what multivariate analysis is and explains that multivariate datasets contain multiple variables and can include mixed data types. Common MVA methods are discussed, including hierarchical cluster analysis (HCA) and partition clustering for exploratory analysis. HCA involves calculating distances between variables, building a tree to visualize relationships, and can identify potential subgroups. Partition clustering assigns variables to discrete clusters. The document demonstrates HCA and partition clustering on gene expression data to explore patterns among patients and genes.
CLUSTERING DICHOTOMOUS DATA FOR HEALTH CAREijistjournal
Dichotomous data is a type of categorical data, which is binary with categories zero and one. Health care data is one of the heavily used categorical data. Binary data are the simplest form of data used for heath care databases in which close ended questions can be used; it is very efficient based on computational efficiency and memory capacity to represent categorical type data. Clustering health care or medical data is very tedious due to its complex data representation models, high dimensionality and data sparsity. In this paper, clustering is performed after transforming the dichotomous data into real by wiener transformation. The proposed algorithm can be usable for determining the correlation of the health disorders and symptoms observed in large medical and health binary databases. Computational results show that the clustering based on Wiener transformation is very efficient in terms of objectivity and subjectivity.
The document discusses data analytics solutions involving machine learning and statistical modeling. It proposes splitting the solution into two parts: 1) applying algorithm techniques and statistical tests to data, and 2) making data-driven decisions using insights, metrics, and innovations. It then provides more details on machine learning techniques like training/testing data sets, and determining the optimal number of neurons in neural networks. Statistical modeling techniques like logistic regression, decision trees, and neural networks are recommended. The document emphasizes comparing different model results to identify ways to improve performance.
1) The document discusses using various machine learning and predictive analytics techniques like random forests, support vector machines, neural networks, and network analysis to analyze healthcare claims data and detect anomalies and fraud.
2) Over 87 million claim lines from 17,000 providers and 1.6 million members are analyzed monthly using 70 measures of claiming behavior and 6 algorithms. Small clusters of providers with high-risk claiming patterns are identified.
3) Random forest classification and support vector machines are used to predict which investigations will yield findings, and random forests provide a measure of variable importance. Network analysis can reveal problematic relationships between providers and members.
Comparative Analysis of Weighted Emphirical Optimization Algorithm and Lazy C...IIRindia
Health care has millions of centric data to discover the essential data is more important. In data mining the discovery of hidden information can be more innovative and useful for much necessity constraint in the field of forecasting, patient’s behavior, executive information system, e-governance the data mining tools and technique play a vital role. In Parkinson health care domain the hidden concept predicts the possibility of likelihood of the disease and also ensures the important feature attribute. The explicit patterns are converted to implicit by applying various algorithms i.e., association, clustering, classification to arrive at the full potential of the medical data. In this research work Parkinson dataset have been used with different classifiers to estimate the accuracy, sensitivity, specificity, kappa and roc characteristics. The proposed weighted empirical optimization algorithm is compared with other classifiers to be efficient in terms of accuracy and other related measures. The proposed model exhibited utmost accuracy of 87.17% with a robust kappa statistics measurement and roc degree indicated the strong stability of the model when compared to other classifiers. The total penalty cost generated by the proposed model is less when compared with the penalty cost of other classifiers in addition to accuracy and other performance measures.
Lead dbs Workshop 2020 Brisbane ProgrammeAndreas Horn
The document describes a 2-day workshop on deep brain stimulation (DBS) neuroimaging techniques using the Lead-DBS software package. The workshop will cover topics like electrode localization, spatial normalization, connectivity mapping, and will provide hands-on sessions for participants to practice techniques on their own data. Attendees should have some experience with MATLAB and Lead-DBS and bring their own laptops with required software installed. The agenda includes sessions on imaging pipelines, connectomics, troubleshooting, and group analysis methods.
- Deep brain stimulation (DBS) of the subthalamic nucleus (STN) in Parkinson's disease patients affects both motor and non-motor functions through interactions with motor, associative, and limbic networks in the basal ganglia.
- Connectomics analysis using fiber tracking from DBS electrode locations can help explain individual variability in behavioral effects of STN-DBS across different cognitive tasks. Specifically, it shows that stimulation of fibers connecting regions like the pre-SMA can predict detriments in stopping behavior.
- Stimulation of prefrontal fibers bypassing the STN that connect to brainstem regions has been linked to worsening of depressive symptoms after surgery through connectomics analysis.
This document discusses linear deformations and basic volumetric imaging concepts relevant to Lead-DBS workflows. It describes how linear (affine) transformations can be used to register different images of the same subject by preserving properties like parallel lines and ratios between points. These transformations map between voxel and world coordinate systems using transformation matrices stored in image headers. Rigid and affine transformations involving translation, rotation, scaling, and shearing are presented for realigning one image to another. Hands-on examples for importing and co-registering images are also mentioned.
This document discusses connectomic deep brain stimulation and summarizes recent findings. It describes a tract commonly seen in STN-DBS and ALIC-DBS that traverses within the anterior limb of the internal capsule (ALIC). While this tract was previously referred to as the medial forebrain bundle (MFB), the document argues it is not the MFB but is similar to the "sl-MFB." Connectivity to medial and lateral prefrontal cortices and a potential hyperdirect pathway from the dorsal anterior cingulate cortex are discussed as being important. The anterior thalamic radiation (ATR) is also potentially linked.
This document summarizes several online tools for neuroimaging and scientific communication. It describes tools for visualizing large datasets like MicroDraw and Neuroglancer. It also outlines tools for collaboration such as Brain Box and Open Neuro Lab. Additionally, it discusses tools for publishing and organizing research like OSF, BioRxiv, Papers, and Figshare. Finally, the document presents tools for communication and project management, including ResearchGate, Github, Slack, and Trello.
Candidate young stellar objects in the S-cluster: Kinematic analysis of a sub...Sérgio Sacani
Context. The observation of several L-band emission sources in the S cluster has led to a rich discussion of their nature. However, a definitive answer to the classification of the dusty objects requires an explanation for the detection of compact Doppler-shifted Brγ emission. The ionized hydrogen in combination with the observation of mid-infrared L-band continuum emission suggests that most of these sources are embedded in a dusty envelope. These embedded sources are part of the S-cluster, and their relationship to the S-stars is still under debate. To date, the question of the origin of these two populations has been vague, although all explanations favor migration processes for the individual cluster members. Aims. This work revisits the S-cluster and its dusty members orbiting the supermassive black hole SgrA* on bound Keplerian orbits from a kinematic perspective. The aim is to explore the Keplerian parameters for patterns that might imply a nonrandom distribution of the sample. Additionally, various analytical aspects are considered to address the nature of the dusty sources. Methods. Based on the photometric analysis, we estimated the individual H−K and K−L colors for the source sample and compared the results to known cluster members. The classification revealed a noticeable contrast between the S-stars and the dusty sources. To fit the flux-density distribution, we utilized the radiative transfer code HYPERION and implemented a young stellar object Class I model. We obtained the position angle from the Keplerian fit results; additionally, we analyzed the distribution of the inclinations and the longitudes of the ascending node. Results. The colors of the dusty sources suggest a stellar nature consistent with the spectral energy distribution in the near and midinfrared domains. Furthermore, the evaporation timescales of dusty and gaseous clumps in the vicinity of SgrA* are much shorter ( 2yr) than the epochs covered by the observations (≈15yr). In addition to the strong evidence for the stellar classification of the D-sources, we also find a clear disk-like pattern following the arrangements of S-stars proposed in the literature. Furthermore, we find a global intrinsic inclination for all dusty sources of 60 ± 20◦, implying a common formation process. Conclusions. The pattern of the dusty sources manifested in the distribution of the position angles, inclinations, and longitudes of the ascending node strongly suggests two different scenarios: the main-sequence stars and the dusty stellar S-cluster sources share a common formation history or migrated with a similar formation channel in the vicinity of SgrA*. Alternatively, the gravitational influence of SgrA* in combination with a massive perturber, such as a putative intermediate mass black hole in the IRS 13 cluster, forces the dusty objects and S-stars to follow a particular orbital arrangement. Key words. stars: black holes– stars: formation– Galaxy: center– galaxies: star formation
Discovery of An Apparent Red, High-Velocity Type Ia Supernova at 𝐳 = 2.9 wi...Sérgio Sacani
We present the JWST discovery of SN 2023adsy, a transient object located in a host galaxy JADES-GS
+
53.13485
−
27.82088
with a host spectroscopic redshift of
2.903
±
0.007
. The transient was identified in deep James Webb Space Telescope (JWST)/NIRCam imaging from the JWST Advanced Deep Extragalactic Survey (JADES) program. Photometric and spectroscopic followup with NIRCam and NIRSpec, respectively, confirm the redshift and yield UV-NIR light-curve, NIR color, and spectroscopic information all consistent with a Type Ia classification. Despite its classification as a likely SN Ia, SN 2023adsy is both fairly red (
�
(
�
−
�
)
∼
0.9
) despite a host galaxy with low-extinction and has a high Ca II velocity (
19
,
000
±
2
,
000
km/s) compared to the general population of SNe Ia. While these characteristics are consistent with some Ca-rich SNe Ia, particularly SN 2016hnk, SN 2023adsy is intrinsically brighter than the low-
�
Ca-rich population. Although such an object is too red for any low-
�
cosmological sample, we apply a fiducial standardization approach to SN 2023adsy and find that the SN 2023adsy luminosity distance measurement is in excellent agreement (
≲
1
�
) with
Λ
CDM. Therefore unlike low-
�
Ca-rich SNe Ia, SN 2023adsy is standardizable and gives no indication that SN Ia standardized luminosities change significantly with redshift. A larger sample of distant SNe Ia is required to determine if SN Ia population characteristics at high-
�
truly diverge from their low-
�
counterparts, and to confirm that standardized luminosities nevertheless remain constant with redshift.
Compositions of iron-meteorite parent bodies constrainthe structure of the pr...Sérgio Sacani
Magmatic iron-meteorite parent bodies are the earliest planetesimals in the Solar System,and they preserve information about conditions and planet-forming processes in thesolar nebula. In this study, we include comprehensive elemental compositions andfractional-crystallization modeling for iron meteorites from the cores of five differenti-ated asteroids from the inner Solar System. Together with previous results of metalliccores from the outer Solar System, we conclude that asteroidal cores from the outerSolar System have smaller sizes, elevated siderophile-element abundances, and simplercrystallization processes than those from the inner Solar System. These differences arerelated to the formation locations of the parent asteroids because the solar protoplane-tary disk varied in redox conditions, elemental distributions, and dynamics at differentheliocentric distances. Using highly siderophile-element data from iron meteorites, wereconstruct the distribution of calcium-aluminum-rich inclusions (CAIs) across theprotoplanetary disk within the first million years of Solar-System history. CAIs, the firstsolids to condense in the Solar System, formed close to the Sun. They were, however,concentrated within the outer disk and depleted within the inner disk. Future modelsof the structure and evolution of the protoplanetary disk should account for this dis-tribution pattern of CAIs.
1. GROUP ANALYSES IN LEAD GROUP
LEAD-DBS WORKSHOP QUEENSLAND| FEB 2020 | Friederike Irmen
2. DBS imaging on a group level (Treu et al. 2020) -
S1 – Walkthrough Tutorial
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
3. Sample
• 51 PD patients (age 60 ± 7.9; 17 female)
• Stimulation parameters
• Regressor: UPDRS-III ON vs. OFF DBS, OFF
Medication:
45.4 ± 23.0% improvement
Is the improvement of motor symptoms related to
electrode location?
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
5. 6.3 mA
andreas.horn@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
1. Define a folder where to store the group analysis “model”
You can switch back and forth between Lead-DBS and Lead Group
This file is auto generated:
- All info stored in here
- more or less portable
project data, i.e. built to share
with colleagues
(restrictions apply)
6. 6.3 mA
You can add patients to your group analysis
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
7. 6.3 mA
You can assign subgroups if needed
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
8. Click group colors and choose preferred colors
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
9. 6.3 mA
Click Review/Edit under Variables and enter values for each patient.
Note: alternatively, variables can be added more easily within Matlab, in the Lead
group file: M.clinical.vars
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
10. Under Stimulation
Parameters, specify active
contacts and amplitudes for
all patients
Select VTA model approach,
add amplitude
and cathodal/anodal contacts.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
11. Select VTA model approach,
add amplitude
and cathodal/anodal contacts.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
12. 6.3 mA
You can enter up to 4 power sources each with
different amplitudes/active contacts
Color of contacts shows which source is
currently being programmed
same color
same color
Jump to next patient and after adding parameters for all patients click save
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
13. Stim params and Group colors buttons turn green if information has been stored in model
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
14. 6.3 mA
Select atlas and “Visualize 3D” the selected patients
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
15. 6.3 mA
You can select/deselect groups/electrodes to visualize
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
16. 2D Visualization of electrodes as point
clouds, colored by group
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
17. 6.3 mA
Select atlas and click “Visualize 2D”
Tune 2D visualization settings
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
18. 2D slices of the most dorsal contacts in the left hemisphere
Light blue and light red colors represent poor vs. good responders, respectively and
dark blue vs. dark red circles depict the two patients with the least and most clinical benefit.
STN is shown in orange.
“Color by regressor”
This box can only be checked if Point-Clouds are selected.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
19. a. Same setup as for Figure 2, now under 3D options
select Point-Clouds and press Visualize 3D
b. Together with the electrode scene, the slice control
figure pops up (bottom left), where the 7T template can
be chosen as background.
3D Visualization of electrodes as point
clouds, colored by group
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
20. 6.3 mA
Select “Highlight Active Contacts” to mark active contacts in red
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
21. 6.3 mA
Now you can specify also in 3D whether to show passive and/or active contacts and whether
to highlight active ones
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
22. Color point-cloud by regressor
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
23. In the function ea_showelectrode.m, the size of the Point-Clouds can be edited in
line 206-212 (e.g. ms=15)
In the viewer: Edit - Current Object Properties can be selected to e.g. the STN
edge color, transparency etc.
Good
responders
Poor
responders
Visualize groups separately
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
24. Map regressor as interpolated point mesh
a. Under Visualization Options check the box “Map
regressor to Coords”
b. Uncheck the box “Show Active Contacts”
c. In 3D options, same as for previous figure and
select “Visualize regressor as: Interpolated point
mesh”
“Visualize regressor as: isosurface”
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
25. Map regressor on VTA
Group
only left hemisphere
Single patient
only left hemisphere
Together with the plot, a window
appears, which allows to change the
threshold, alpha and smoothing of
the regressor
The output, provided for the left
hemisphere, can be found in the
group folder under statvat_results-
models
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
26. Mapping regressors in Lead group
electrode
coordinates as
interpolated
point mesh
point clouds
colored by
clinical
improvement
Visualize
regressor as
isosurface
VTAs of patients
weighted by
clinical
improvement
and thresholded
by a T-value of
10.56;
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
27. The “Target report” feature
• Calculate distance to a target
• Differentiate contacts that lie within the target
structure vs. outside it
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
28. The “Target report” feature
Careful: Since it’s always the distance to the closest voxel of a
target, even if the electrode contact is in the middle of the
target structure, the distance to the shell of the target will be
listed. Thus, a visual check of the contact positions is
recommended!
Distance from each contact to the
closest (!) voxel of the target
structure
indicating whether each contact lies
“within” the target, based on a chosen
threshold.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
29. The “Target report” feature
For each active contact (highlighted in red), the distance from the contact center to the
STN is shown for patients with best (red) and worst (blue) clinical outcomes.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
30. VTA intersections
Plots of the correlations and group comparisons can be generated under
“Explore Stats” – “Analyze Intersections with local structures”.
Depending on the chosen atlas under General settings, subcortical regions
can be chosen. Clinical variables/groups need to be selected, respectively.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
31. VTA intersection with target structure
Correlation between clinical
improvement and intersections
between VTA and the STN.
Patients with best vs.
poorest responses are highlighted.
Patients were arbitrarily median-split into two
groups, based on their percentage
improvement. The group of better responders
(red) showed significantly higher VTA overlap
with the STN.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
32. Here, the regions
available depend on
the selected brain
parcellation.
Correlations with connected stuctures
correlation/group comparison with fibercounts to
connected structures
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
33. Correlations with connected stuctures
Correlation with connectivity to cortical structures. The six structures of
the HMAT parcellation, colored by the R value, resulting from
Spearman’s rank-correlations between fibercounts to these areas and
clinical improvement.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
34. Discriminative Fibers
• Under 3D settings, select
“Show discriminative fibers”
• Select whether
positive/negative/both fibers
should be visualized; select
threshold; you can also switch
to the E-field model
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
35. Discriminative Fibers
• “Visualize 3D” and the fibers
will be selected, which may
take a while when executed for
the first time.
• When finished, a control
window pops up, which allows
you to change the thresholds.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS
36. 6.3 mA
Tracts that are positively associated with clinical
improvement are colored by their T-value
Discriminative Fibers
In the toolbar you’ll find “Add objects”, where you can add regions of
interest to the figure, like the HMAT parcellation in blue.
friederike.irmen@charite.deLead-DBS Workshop
Group analyses in Lead-DBS