AWS Machine Learning abstracts a lot of the complexity of a machine learning solution (e.g. cross-validation, training data set management, algorithm selection, F1-score computation) making it easy to train and deploy machine learning models.
2. Workshop Objectives
Identify data sets and problems where machine learning technologies can be applied
Identify features and target variables in data sets
Refine a training data set to use when build machine learning models
Train machine learning models using AWS Machine Learning
Evaluate machine learning models using AWS Machine Learning
Compute batch predictions using AWS Machine Learning
3. Solving Business Problems with Amazon
Machine Learning
● Supervised learning: learning from data that has been
labeled with the actual answer
● You have existing examples of actual answers
● Acute inflammations in patients
○ You can not code the rules
○ You can not scale
4. Classification, Regression
Examples of binary classification problems:
● Is this patient sick?
Examples of multiclass classification
problems:
● Given possible treatments, which will
succeed?
Examples of regression classification
problems:
● How many days pass before a
chronically ill patient returns?
5. Creating a Data Source
● The target – The answer that you want to predict
● Variables/features – These are attributes of the example
that can be used to identify patterns to predict the target
https://s3.amazonaws.com/mikeghen/acute-inflammations.csv
J.Czerniak, H.Zarzycki, Application of rough sets in the presumptive diagnosis of urinary system diseases, Artifical Inteligence and Security in
Computing Systems, ACS'2002 9th International Conference Proceedings, Kluwer Academic Publishers,2003, pp. 41-51
6. Training and Validation of a ML model
● Splits the training datasource into two sections
● Trains the ML model on the section that contains
70% of the input data
● Evaluates the model using the remaining 30% of the
input data
8. Advanced Topics
● Bootstrapping and Boosting
● Deploying models
○ Building ML Models using custom code
○ Creating an endpoint for real-time predictions
● The importance of evaluation
○ FOREX experiences