Predictive Web Analytics is aimed at understanding behavioural patterns of users of various web-based applications: e-commerce, ubiquitous and mobile computing, and computational advertising. Within these applications business decisions often rely on two types of predictions: an overall or particular user segment demand predictions and individualised recommendations for visitors. Visitor behaviour is inherently sensitive to the context, which can be de ned as a collection of external factors. Context-awareness allows integrating external explanatory information into the learning process and adapting user behaviour accordingly. The importance of context-awareness has been recognised by researchers and practitioners in many disciplines, including recommendation systems, information retrieval, personalization, data mining, and marketing. We focus on studying ways of context discovery and its integration into predictive analytics.
Predicting Current User Intent with Contextual Markov ModelsJulia Kiseleva
Abstract—In many web information systems like e-shops and information portals predictive modeling is used to understand user intentions based on their browsing behavior. User behavior is inherently sensitive to various contexts. Identifying such relevant contexts can help to improve the prediction performance. In this work, we propose a formal approach in which the context
discovery process is defined as an optimization problem. For simplicity we assume a concrete yet generic scenario in which context is considered to be a secondary label of an instance that is either known from the available contextual attribute (e.g. user location) or can be induced from the training data (e.g. novice vs. expert user). In an ideal case, the objective function of the optimization problem has an analytical form enabling us
to design a context discovery algorithm solving the optimization problem directly. An example with Markov models, a typical approach for modeling user browsing behavior, shows that the derived analytical form of the optimization problem provides us with useful mathematical insights of the problem. Experiments with a real-world use-case show that we can discover useful contexts allowing us to significantly improve the prediction of
user intentions with contextual Markov models.
Scalable Exploration of Relevance Prospects to Support Decision MakingKatrien Verbert
Presented at IntRS 2016 - Interfaces and Human Decision Making for Recommender Systems, workshop at RecSys 2016
Citation: Verbert, K., Seipp, K., He, C., Parra, D., Wongchokprasitti, C., & Brusilovsky, P. (2016). Scalable Exploration of Relevance Prospects to Support Decision Making. Proceedings of the Joint Workshop on Interfaces and Human Decision Making for Recommender Systems co-located with ACM Conference on Recommender Systems (RecSys 2016), Boston, MA, USA, September 16, 2016.
This slide was presented in International the 2015 Conference on Education Research.
I aggregated several my other partial slides and reports to describe adaptive learning model pertaining to concept of learning analytics as well as LOD for curriculum standards and digital resources. There is short introduction to the project of ISO/IEC 20748 Learning analytics interoperability - Part 1: Reference model.
Imagining the Future through Social Media as a Tool for Social Innovation (E...Mario Guillo
F212.org is a virtual think tank of university students interested in sharing ideas on how to face main future challenges. It describes the results of a comparative study about the images of the future found among young students from Haaga Helia University of Applied Science (Finland) Tamkang University (Taiwan); and University of Alicante (Spain).
FLUX·3D - Forward Looking User eXperienceMario Guillo
FLUX·3D is a tool for making possible to the people participating actively -and in a sustainable way- in the design of products, services and processes. This tool makes possible to approach a complex problem (such as the evaluation -in absolute and relative terms- of an innovation/prototype) in a simple and systematic way, which can be very helpful for making decisions (and defining future strategies) and/or improving the development of an innovation. FLUX·3D has been designed and developed by FUTURLAB - University of Alicante (Spain) together with Aalto University (Finland).
Predicting Current User Intent with Contextual Markov ModelsJulia Kiseleva
Abstract—In many web information systems like e-shops and information portals predictive modeling is used to understand user intentions based on their browsing behavior. User behavior is inherently sensitive to various contexts. Identifying such relevant contexts can help to improve the prediction performance. In this work, we propose a formal approach in which the context
discovery process is defined as an optimization problem. For simplicity we assume a concrete yet generic scenario in which context is considered to be a secondary label of an instance that is either known from the available contextual attribute (e.g. user location) or can be induced from the training data (e.g. novice vs. expert user). In an ideal case, the objective function of the optimization problem has an analytical form enabling us
to design a context discovery algorithm solving the optimization problem directly. An example with Markov models, a typical approach for modeling user browsing behavior, shows that the derived analytical form of the optimization problem provides us with useful mathematical insights of the problem. Experiments with a real-world use-case show that we can discover useful contexts allowing us to significantly improve the prediction of
user intentions with contextual Markov models.
Scalable Exploration of Relevance Prospects to Support Decision MakingKatrien Verbert
Presented at IntRS 2016 - Interfaces and Human Decision Making for Recommender Systems, workshop at RecSys 2016
Citation: Verbert, K., Seipp, K., He, C., Parra, D., Wongchokprasitti, C., & Brusilovsky, P. (2016). Scalable Exploration of Relevance Prospects to Support Decision Making. Proceedings of the Joint Workshop on Interfaces and Human Decision Making for Recommender Systems co-located with ACM Conference on Recommender Systems (RecSys 2016), Boston, MA, USA, September 16, 2016.
This slide was presented in International the 2015 Conference on Education Research.
I aggregated several my other partial slides and reports to describe adaptive learning model pertaining to concept of learning analytics as well as LOD for curriculum standards and digital resources. There is short introduction to the project of ISO/IEC 20748 Learning analytics interoperability - Part 1: Reference model.
Imagining the Future through Social Media as a Tool for Social Innovation (E...Mario Guillo
F212.org is a virtual think tank of university students interested in sharing ideas on how to face main future challenges. It describes the results of a comparative study about the images of the future found among young students from Haaga Helia University of Applied Science (Finland) Tamkang University (Taiwan); and University of Alicante (Spain).
FLUX·3D - Forward Looking User eXperienceMario Guillo
FLUX·3D is a tool for making possible to the people participating actively -and in a sustainable way- in the design of products, services and processes. This tool makes possible to approach a complex problem (such as the evaluation -in absolute and relative terms- of an innovation/prototype) in a simple and systematic way, which can be very helpful for making decisions (and defining future strategies) and/or improving the development of an innovation. FLUX·3D has been designed and developed by FUTURLAB - University of Alicante (Spain) together with Aalto University (Finland).
Agents vs Users: Visual Recommendation of Research Talks with Multiple Dimens...Katrien Verbert
Published in ACM TiiS: Verbert, K., Parra, D., & Brusilovsky, P. (2016). Agents Vs. Users: Visual Recommendation of Research Talks with Multiple Dimension of Relevance. ACM Transactions on Interactive Intelligent Systems (TiiS), 6(2), 11.
Presented at IUI 2017
Behavioural Modelling Outcomes prediction using Casual FactorsIJMER
Generating models from large data sets—and deter-mining which subsets of data to
mine—is becoming increasingly automated. However choosing what data to collect in the first place
requires human intuition or experience, usually supplied by a domain expert. This paper describes a
new approach to machine science which demonstrates for the first time that non-domain experts can
collectively formulate features, and provide values for those features such that they are predictive of
some behavioral outcome of interest. This was accomplished by building a web platform in which
human groups interact to both respond to questions likely to help predict a behavioral outcome and
pose new questions to their peers. This results in a dynamically-growing online survey, but the result
of this cooperative behavior also leads to models that can predict user's outcomes based on their
responses to the user-generated survey questions. Here we describe two web-based experiments that
instantiate this approach: the first site led to models that can predict users' monthly electric energy
consumption; the other led to models that can predict users' body mass index. As exponential
increases in content are often observed in successful online collaborative communities, the proposed
methodology may, in the future, lead to similar exponential rises in discovery and insight into the
causal factors of behavioral outcomes
A recommender system(RS) is an information filtering system that recommends the related suggestions as per the end users requirement. Applications of RS include recommendation of movies, music, serials, books, documents, websites, tourist places etc.
Benefits of RS: RSs are beneficial to both service providers and to the users. RSs reduce transaction costs of finding and selecting items.& RSs help in decision making. The proposed work DEMOGRAPHY BASED HYBRID SYSTEM FOR MOVIE RECOMMENDATIONS highlights the combination of collaborative, content based & demographic filtering to recommend movies to the end user. The model uses SVD++ technique available in Surprise Python library and achieves the MSE of 0.92 which is comparatively less than the other techniques.
Strategies for Practical Active Learning, Robert MunroRobert Munro
In many real-world Machine Learning applications, you need to continually update your models with new training data to improve and maintain accuracy as your model is applied. However, it is often difficult to decide what new data needs to be labeled for training, and what is the best workflow and interfaces for labeling. This training will focus on how you can use Active Learning to improve your training data at scale with common Deep Learning frameworks. At the end of this session, you will understand several Active Learning strategies. We will use the example of applying Active Learning to the ImageNet data set using the TensorFlow Deep Learning framework.
Using Large-Scale LMS Data Portal Data to Improve Teaching and Learning (at K...Shalin Hai-Jew
With any learning management system, a byproduct of its function is data, which may be analyzed to improve awareness, decision-making, and actions. At Kansas State University, its Canvas LMS instance recently made available its cumulative data from its first use in 2013. These flat files open a window to how the university is harnessing its LMS, with some macro-level insights that may suggest some areas to improve teaching and learning. This session describes some approaches to informatizing this empirical “big data” with some basic approaches: reviewing the data dictionary, extracting basic descriptions of the respective data sets, conducting time-based comparisons, surfacing testable hypotheses from data inferences, and conducting other data explorations. This introduces initial data analysis work only, but this does not preclude front-end analysis of courses at the micro level, relational database queries of the data, and other potential follow-on work.
Analysis of Online Product Purchase and Predicting Items for Co-purchase - IC...Sohom Ghosh
In recent years, online market places have become popular
among the buyers. During this course of time, not only have they sustained the business model but also generated large amount of profit,
turning it into a lucrative business model. In this paper, we take a look
at a temporal dataset from one of the most successful online businesses
to analyze the nature of the buying patterns of the users. Arguably,
the most important purchase characteristic of such networks is follow-up
purchase by a buyer, otherwise known as a co-purchase. In this paper,
we also analyze the co-purchase patterns to build a knowledge-base to
recommend potential co-purchase items for every item.
In this presentation we illustrate a novel context-aware mobile recommender system for places of interest (POIs). Unlike existing systems, which learn users’ preferences solely from their past ratings, it considers also their personality - using the Five Factor Model. Personality is acquired by asking users to complete a brief and entertaining questionnaire as part of the registration process, and is then exploited in: (1) an active learning module that actively acquires ratings-in-context for POIs that users are likely to have experienced, hence reducing the stress and annoyance to rate (or skip rating) items that the users don’t know; and (2) in the recommendation model that builds up on matrix factorization and therefore can be trained even if the users haven’t rated any items yet.
Discovering Common Motifs in Cursor Movement DataYandex
Mouse cursor movements can provide valuable information on how users interact and engage with web documents. This interaction data is far richer than traditional click data, and can be used to improve evaluation and presentation of web information systems. Unfortunately, the diversity and complexity inherent in this interaction data make it more difficult to capture salient behavior characteristics through traditional feature engineering. To address this problem, we introduce a novel approach of automatically discovering frequent subsequences, or motifs, in mouse cursor movement data. In order to scale our approach to realistic datasets, we introduce novel optimizations for motif discovery, specifically designed for mining cursor movement data. We show that by encoding the motifs discovered from thousands of real web search sessions as features, enables significant improvements in important web search tasks. These results, complemented with visualization and qualitative analysis, demonstrate that our approach is able to automatically capture key characteristics of mouse cursor movement behavior, providing a valuable new tool for online user behavior analysis. In addition to the application of motifs to web mining, we demonstrate that similar technique can be successfully applied in medical domain for the task of predicting future decline of memory function and subsequent development of the Alzheimer Disease.
Graphs for Recommendation Engines: Looking beyond Social, Retail, and MediaNeo4j
We’re all familiar with recommendations in a number of different areas of our lives. Recommendations for social media connections, e-commerce products, or streaming media content are ubiquitous.
Perhaps less well known are applications for recommendations in different contexts, like education, HR, fraud, business process management, or offender rehabilitation.
In this webinar, we will discuss some of these recommendations use cases in more detail, and look at how graph data can be used to model each domain and power a recommendations engine. We’ll also see an example use case demonstrated using Neo4j.
Agents vs Users: Visual Recommendation of Research Talks with Multiple Dimens...Katrien Verbert
Published in ACM TiiS: Verbert, K., Parra, D., & Brusilovsky, P. (2016). Agents Vs. Users: Visual Recommendation of Research Talks with Multiple Dimension of Relevance. ACM Transactions on Interactive Intelligent Systems (TiiS), 6(2), 11.
Presented at IUI 2017
Behavioural Modelling Outcomes prediction using Casual FactorsIJMER
Generating models from large data sets—and deter-mining which subsets of data to
mine—is becoming increasingly automated. However choosing what data to collect in the first place
requires human intuition or experience, usually supplied by a domain expert. This paper describes a
new approach to machine science which demonstrates for the first time that non-domain experts can
collectively formulate features, and provide values for those features such that they are predictive of
some behavioral outcome of interest. This was accomplished by building a web platform in which
human groups interact to both respond to questions likely to help predict a behavioral outcome and
pose new questions to their peers. This results in a dynamically-growing online survey, but the result
of this cooperative behavior also leads to models that can predict user's outcomes based on their
responses to the user-generated survey questions. Here we describe two web-based experiments that
instantiate this approach: the first site led to models that can predict users' monthly electric energy
consumption; the other led to models that can predict users' body mass index. As exponential
increases in content are often observed in successful online collaborative communities, the proposed
methodology may, in the future, lead to similar exponential rises in discovery and insight into the
causal factors of behavioral outcomes
A recommender system(RS) is an information filtering system that recommends the related suggestions as per the end users requirement. Applications of RS include recommendation of movies, music, serials, books, documents, websites, tourist places etc.
Benefits of RS: RSs are beneficial to both service providers and to the users. RSs reduce transaction costs of finding and selecting items.& RSs help in decision making. The proposed work DEMOGRAPHY BASED HYBRID SYSTEM FOR MOVIE RECOMMENDATIONS highlights the combination of collaborative, content based & demographic filtering to recommend movies to the end user. The model uses SVD++ technique available in Surprise Python library and achieves the MSE of 0.92 which is comparatively less than the other techniques.
Strategies for Practical Active Learning, Robert MunroRobert Munro
In many real-world Machine Learning applications, you need to continually update your models with new training data to improve and maintain accuracy as your model is applied. However, it is often difficult to decide what new data needs to be labeled for training, and what is the best workflow and interfaces for labeling. This training will focus on how you can use Active Learning to improve your training data at scale with common Deep Learning frameworks. At the end of this session, you will understand several Active Learning strategies. We will use the example of applying Active Learning to the ImageNet data set using the TensorFlow Deep Learning framework.
Using Large-Scale LMS Data Portal Data to Improve Teaching and Learning (at K...Shalin Hai-Jew
With any learning management system, a byproduct of its function is data, which may be analyzed to improve awareness, decision-making, and actions. At Kansas State University, its Canvas LMS instance recently made available its cumulative data from its first use in 2013. These flat files open a window to how the university is harnessing its LMS, with some macro-level insights that may suggest some areas to improve teaching and learning. This session describes some approaches to informatizing this empirical “big data” with some basic approaches: reviewing the data dictionary, extracting basic descriptions of the respective data sets, conducting time-based comparisons, surfacing testable hypotheses from data inferences, and conducting other data explorations. This introduces initial data analysis work only, but this does not preclude front-end analysis of courses at the micro level, relational database queries of the data, and other potential follow-on work.
Analysis of Online Product Purchase and Predicting Items for Co-purchase - IC...Sohom Ghosh
In recent years, online market places have become popular
among the buyers. During this course of time, not only have they sustained the business model but also generated large amount of profit,
turning it into a lucrative business model. In this paper, we take a look
at a temporal dataset from one of the most successful online businesses
to analyze the nature of the buying patterns of the users. Arguably,
the most important purchase characteristic of such networks is follow-up
purchase by a buyer, otherwise known as a co-purchase. In this paper,
we also analyze the co-purchase patterns to build a knowledge-base to
recommend potential co-purchase items for every item.
In this presentation we illustrate a novel context-aware mobile recommender system for places of interest (POIs). Unlike existing systems, which learn users’ preferences solely from their past ratings, it considers also their personality - using the Five Factor Model. Personality is acquired by asking users to complete a brief and entertaining questionnaire as part of the registration process, and is then exploited in: (1) an active learning module that actively acquires ratings-in-context for POIs that users are likely to have experienced, hence reducing the stress and annoyance to rate (or skip rating) items that the users don’t know; and (2) in the recommendation model that builds up on matrix factorization and therefore can be trained even if the users haven’t rated any items yet.
Discovering Common Motifs in Cursor Movement DataYandex
Mouse cursor movements can provide valuable information on how users interact and engage with web documents. This interaction data is far richer than traditional click data, and can be used to improve evaluation and presentation of web information systems. Unfortunately, the diversity and complexity inherent in this interaction data make it more difficult to capture salient behavior characteristics through traditional feature engineering. To address this problem, we introduce a novel approach of automatically discovering frequent subsequences, or motifs, in mouse cursor movement data. In order to scale our approach to realistic datasets, we introduce novel optimizations for motif discovery, specifically designed for mining cursor movement data. We show that by encoding the motifs discovered from thousands of real web search sessions as features, enables significant improvements in important web search tasks. These results, complemented with visualization and qualitative analysis, demonstrate that our approach is able to automatically capture key characteristics of mouse cursor movement behavior, providing a valuable new tool for online user behavior analysis. In addition to the application of motifs to web mining, we demonstrate that similar technique can be successfully applied in medical domain for the task of predicting future decline of memory function and subsequent development of the Alzheimer Disease.
Graphs for Recommendation Engines: Looking beyond Social, Retail, and MediaNeo4j
We’re all familiar with recommendations in a number of different areas of our lives. Recommendations for social media connections, e-commerce products, or streaming media content are ubiquitous.
Perhaps less well known are applications for recommendations in different contexts, like education, HR, fraud, business process management, or offender rehabilitation.
In this webinar, we will discuss some of these recommendations use cases in more detail, and look at how graph data can be used to model each domain and power a recommendations engine. We’ll also see an example use case demonstrated using Neo4j.
How Universities Use Big Data to Transform EducationHortonworks
Student performance data is increasingly being captured as part of software-based and online classroom exercises and testing. This data can be augmented with behavioral data captured from sources such as social media, student-professor meeting notes, blogs, student surveys, and so forth to discover new insights to improve student learning. The results transcend traditional IT departments to focus on issues like retention, research, and the delivery of content and courses through new modalities.
Hortonworks is partnering with Microsoft to show you how the Hortonworks Data Platform (HDP) running on the Microsoft stack enables you to develop a “single view of a student”.
Understanding and predicting behavior for each individual customer has always been the ultimate dream for all digital companies. Combining machine learning and big data processing has finally made that dream a reality. In this webcast, you'll learn about the behavior based algorithms Insights uses to predict customer behavior.
Listen to the podcast version here: http://bit.ly/1EYkSIH
View the webcast on Youtube: https://youtu.be/sidTdUkacHw
Personalized Search-Building a prototype to infer the user's interestTom Burgmans
In the world of Search, understanding the intend of the user is often seen as the holy grail. When a user performs multiple search and click actions while having a conversation with the search engine, then this behavior reveals a piece of her/his interest. A search engine that is aware of the user’s interest is able to add a personal layer in its responses and this could add a new dimension of accuracy and value to a search implementation. But what technology does it take to build it? What data is needed? How well does it really work? This presentation describes the journey to find a practical implementation of a recommendation engine. It answers all the questions above and more. We’ll guide you through the lessons learned while creating an engine that generates potentially interesting items for the user based on collaborative filtering and anomaly detection. We’ll demonstrate a prototype where even a minimal set of user actions could lead to a personalized search experience.
An institutional perspective on analytics that focusses on a particular tool developed using an agile methodology to visualise learner behaviours in MOOCs via Sankey diagrams.
Predicting User Satisfaction with Intelligent AssistantsJulia Kiseleva
There is a rapid growth in the use of voice-controlled intelligent
personal assistants on mobile devices, such as Microsoft’s Cortana,
Google Now, and Apple’s Siri.
They significantly change the way users interact with search systems,
not only because of the voice control use and touch gestures,
but also due to the dialogue-style nature of the interactions and their
ability to preserve context across different queries. Predicting success
and failure of such search dialogues is a new problem, and
an important one for evaluating and further improving intelligent
assistants. While clicks in web search have been extensively used
to infer user satisfaction, their significance in search dialogues is
lower due to the partial replacement of clicks with voice control,
direct and voice answers, and touch gestures.
In this paper, we propose an automatic method to predict user
satisfaction with intelligent assistants that exploits all the interaction
signals, including voice commands and physical touch gestures
on the device.
First, we conduct an extensive user study to measure user satisfaction
with intelligent assistants, and simultaneously record all
user interactions. Second, we show that the dialogue style of interaction
makes it necessary to evaluate the user experience at the
overall task level as opposed to the query level. Third, we train a
model to predict user satisfaction, and find that interaction signals
that capture the user reading patterns have a high impact: when including
all available interaction signals, we are able to improve the
prediction accuracy of user satisfaction from 71% to 81% over a
baseline that utilizes only click and query features.
Using Contextual Information to Understand Searching and Browsing BehaviorJulia Kiseleva
Julia Kiseleva's slides for PhD defense on June 13 2016.
The thesis is available by the following link -- https://www.researchgate.net/publication/303285745_Using_Contextual_Information_to_Understand_Searching_and_Browsing_Behavior
Detecting Good Abandonment in Mobile SearchJulia Kiseleva
Web search queries for which there are no clicks are referred to as abandoned queries and are usually considered
as leading to user dissatisfaction. However, there are many
cases where a user may not click on any search result page
(SERP) but still be satised. This scenario is referred to
as good abandonment and presents a challenge for most ap-
proaches measuring search satisfaction, which are usually
based on clicks and dwell time. The problem is exacerbated
further on mobile devices where search providers try to in-
crease the likelihood of users being satised directly by the
SERP. This paper proposes a solution to this problem us-
ing gesture interactions, such as reading times and touch
actions, as signals for dierentiating between good and bad
abandonment. These signals go beyond clicks and charac-
terize user behavior in cases where clicks are not needed to
achieve satisfaction. We study different good abandonment
scenarios and investigate the dierent elements on a SERP
that may lead to good abandonment. We also present an
analysis of the correlation between user gesture features and
satisfaction. Finally, we use this analysis to build models to
automatically identify good abandonment in mobile search
achieving an accuracy of 75%, which is significantly better
than considering query and session signals alone. Our fundings have implications for the study and application of user
satisfaction in search systems.
Understanding User Satisfaction with Intelligent AssistantsJulia Kiseleva
Voice-controlled intelligent personal assistants, such as Cortana,
Google Now, Siri and Alexa, are increasingly becoming a part of
users’ daily lives, especially on mobile devices. They introduce
a significant change in information access, not only by introducing
voice control and touch gestures but also by enabling dialogues
where the context is preserved. This raises the need for evaluation
of their effectiveness in assisting users with their tasks. However,
in order to understand which type of user interactions reflect different
degrees of user satisfaction we need explicit judgements. In this
paper, we describe a user study that was designed to measure user
satisfaction over a range of typical scenarios of use: controlling a
device, web search, and structured search dialogue. Using this data,
we study how user satisfaction varied with different usage scenarios
and what signals can be used for modeling satisfaction in the
different scenarios. We find that the notion of satisfaction varies
across different scenarios, and show that, in some scenarios (e.g.
making a phone call), task completion is very important while for
others (e.g. planning a night out), the amount of effort spent is key.
We also study how the nature and complexity of the task at hand
affects user satisfaction, and find that preserving the conversation
context is essential and that overall task-level satisfaction cannot
be reduced to query-level satisfaction alone. Finally, we shed light
on the relative effectiveness and usefulness of voice-controlled intelligent
agents, explaining their increasing popularity and uptake
relative to the traditional query-response interaction.
Behavioral Dynamics from the SERP’s Perspective: What are Failed SERPs and Ho...Julia Kiseleva
Web search is always in a state of flux: queries, their intent, and the
most relevant content are changing over time, in predictable and unpredictable
ways. Modern search technology has made great strides
in keeping up to pace with these changes, but there remain cases of
failure where the organic search results on the search engine result
page (SERP) are outdated, and no relevant result is displayed.
Failing SERPs due to temporal drift are one of the greatest frustrations
of web searchers, leading to search abandonment or even
search engine switch. Detecting failed SERPs timely and providing
access to the desired out-of-SERP results has huge potential to improve
user satisfaction. Our main findings are threefold: First, we
refine the conceptual model of behavioral dynamics on the web by
including the SERP and defining (un)successful SERPs in terms of
observable behavior. Second, we analyse typical patterns of temporal
change and propose models to predict query drift beyond the
current SERP, and ways to adapt the SERP to include the desired
results. Third, we conduct extensive experiments on real world
search engine traffic demonstrating the viability of our approach.
Our analysis of behavioral dynamics at the SERP level gives new
insight in one of the primary causes of search failure due to temporal
query intent drifts. Our overall conclusion is that the most
detrimental cases in terms of (lack of) user satisfaction lead to the
largest changes in information seeking behavior, and hence to observable
changes in behavior we can exploit to detect failure, and
moreover not only detect them but also resolve them.
Where to Go on Your Next Trip? Optimizing Travel Destinations Based on User P...Julia Kiseleva
Recommendation based on user preferences is a common
task for e-commerce websites. New recommendation algorithms
are often evaluated by offline comparison to baseline
algorithms such as recommending random or the most
popular items. Here, we investigate how these algorithms
themselves perform and compare to the operational production
system in large scale online experiments in a real-world
application. Specifically, we focus on recommending travel
destinations at Booking.com, a major online travel site, to
users searching for their preferred vacation activities. To
build ranking models we use multi-criteria rating data provided
by previous users after their stay at a destination. We
implement three methods and compare them to the current
baseline in Booking.com: random, most popular, and Naive
Bayes. Our general conclusion is that, in an online A/B test
with live users, our Naive-Bayes based ranker increased user
engagement significantly over the current online system
Chatty Kathy - UNC Bootcamp Final Project Presentation - Final Version - 5.23...John Andrews
SlideShare Description for "Chatty Kathy - UNC Bootcamp Final Project Presentation"
Title: Chatty Kathy: Enhancing Physical Activity Among Older Adults
Description:
Discover how Chatty Kathy, an innovative project developed at the UNC Bootcamp, aims to tackle the challenge of low physical activity among older adults. Our AI-driven solution uses peer interaction to boost and sustain exercise levels, significantly improving health outcomes. This presentation covers our problem statement, the rationale behind Chatty Kathy, synthetic data and persona creation, model performance metrics, a visual demonstration of the project, and potential future developments. Join us for an insightful Q&A session to explore the potential of this groundbreaking project.
Project Team: Jay Requarth, Jana Avery, John Andrews, Dr. Dick Davis II, Nee Buntoum, Nam Yeongjin & Mat Nicholas
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Explore our comprehensive data analysis project presentation on predicting product ad campaign performance. Learn how data-driven insights can optimize your marketing strategies and enhance campaign effectiveness. Perfect for professionals and students looking to understand the power of data analysis in advertising. for more details visit: https://bostoninstituteofanalytics.org/data-science-and-artificial-intelligence/
1. Context Mining and Integration
into Web Predictive Analytics
Julia Kiseleva
2. Outline
• What is Web Predictive Analytics
• Context-Aware Predictive Analytics
framework
• Datasets
• User Intent Modeling
• Contextual Markov Models
• Discovering Change in User intent
• Conclusions and Further and Ongoing
work
8/21/2014 2
5. User Intent Modeling:
What?
• Next action prediction
o Click prediction on display advertising
o Drop out prediction
o User Trail prediction
• Information need prediction
o Navigational vs. explorative vs. purchase
o Changes in user intent
• Type of product wanted
o Personalization based on context
o Personalization based on changed context
8/21/2014 5
6. User Intent Modeling:
Why?
• To understand users and website usage
o redesign website
o diversified search
o search recommendations
• To better use advertisement budget
o When serve ads?
o What type of ads to serve?
o brand awareness CPM or convergence CPC
• To manipulate user – worth giving a promotion?
o personalize with intent of converging to a desired
action
o personalized suggestions based on user context
8/21/2014 6
7. Web Predictive Analytics
• Web predictive analytics -
aims to predict individual and aggregated
characteristics indicating visitor behavior for purposes
of understanding and optimizing web usage
• Application:
o Search engines
o Recommender System
o Computational Advertisement
• Predictive web analytics tasks:
o Online shop’s recommendations
o Users’ next action prediction
o Users’ intention predicting
o Personalized search result page 8/21/2014 7
8. User Intent Modeling:
How?
Model L
Users web log
Historical
data
labels
label?
1. training
3. application
X
y
X'
labels
Testing
data
2. testing
Training:
y = L (X)
Application:
use L for an unseen
data
y' = L (X')
Formulations:
① Classification
② Regression
③ Clustering
④ Scoring
8/21/2014 8
9. Type of Context
• User Context
o User Preferences
o User profiles
o Usage of user history
• Document/Product Context
o Meta-data
o Content features
• Task Context
o Current activity
o Location and etc.
• Social Context
o Leveraging the social graph
8/21/2014 9
10. Example of Context:
in Diagnostics
• Not predictive alone but a subset of features with
the contextual attribute(s) becomes (much) more
predictive
Time of
the day
context
No context
8/21/2014 10
11. Example of Context:
in Marketing
P(Purchase|gender=“male”)=P(Purchase|gender=“female”)
ModelMale~f(relevance); ModelFemale~f(perceived
value);
Gender
context
Male
Female
buy
relevance
buydon’t
don’t
No context
buy
relevance
don’t
gender
8/21/2014 11
12. History of context
definition and discovery
Context Year
Location 1992
Taxonomy of explicit context 1999
Predictive features vs. contextual 2002
Hidden context: (clustering, mixture
models)
2004
Contextual bandits 2007
History of previous interaction 2008
Independence of predicted class 2011
Two level prediction model 2012
Focus on Context Discovery 2012 -
Timeline
8/21/2014 12
13. Taxonomy for explicit
Context
Human Factors
Physical
Environment
Factors
User
Characteristics
Social
Environment
Intent
Conditions
Infrastructure
Location
*Weather
*Light
*Acceleration
*Audio
*…
*Temperature
*Humidity
*…
8/21/2014 13
14. Strategies for Context
Discovery
Definitions/pr
operties/utiliti
es
[Un] [Semi]
Supervised
methods
How to define
context
Context mining:
how to discover context
• Contextual features
• Contextual categories
• Features not predictive
alone, but increasing
predictive power of other
features
• Descriptors explaining a
significant group of
instances having some
distinct behaviour
• Subgroup discovery
• AntiLDA
• Uplift modeling
• Actionable attributes
8/21/2014 14
15. Predictive
model(s)
Predictions
Training
data
Context Integration
Output correction
if (context == “spring”)
select
instances(“spring”)
if (context == “spring”)
select models (“spring”)
if (context == “spring”)
score += 0.1*score
Instance set selection
Feature set selection
Feature set expansion Model selection/weighting
Model adjustment
8/21/2014 15
Strategies for Context
Integration
18. Context-Aware Systems
Context definition
Context Integration Method
Application
Context-aware system
Recommendation
systems
Computational
Advertisement
Information
Retrieval
Normalization
Expansion
Classifier Selection
Classification Adjustment
Weighting
Domain Expert
Clustering
Contextual feature identification
8/21/2014 18
19. Research Questions
1. How to define the context (form and maintain
contextual categories) in web analytics?
2. How to connect context with the prediction
process in predictive web analytics?
3. How to integrate change detection mechanisms
into the prediction process in web analytics?
4. How to ensure integration and feedback
mechanisms between change detection and
context awareness mechanisms?
5. What should a reference architecture allowing to
plug in new context aware prediction techniques
for a collection of web analytics tasks look like?
8/21/2014 19
25. Contextual Partitioning
• Approaches to create local models:
o Horizontal partitions
Users
from
Europe
Users
from
South
America
Session 1 Search Refine
Search
Click on
Banner
Product
View
Pay
ment
Session 3 Product
View
Payment
Session 3 Search Refine
Search
Refine
Search
Click on
Banner
Session 4 Search Refine
Search
Click on
Banner
Product
View
Pay
ment
Session 5 Product
View
Click on
Banner
Search
8/21/2014 25
26. Motivation for Contextual Markov
Models
Useful Contexts:
E[M] < pc1*E[Mc1] + pc2*E[Mc2]
Why should it help?
Explicit contexts (user location)
Implicit contexts (inferred from clickstream)
8/21/2014 26
30. Temporal Context-
Awareness
M1
M2
Mk
G H
Temporal Context-Awareness:
(G,H,ti)
……..
t2
t1
tl-1
t3
tl
C2
C1
Ck
G
G
H
H
a1
a2
a3
al-1
al
WebSessionS
Contextual
features
Contextual
Categories
C2
Individual
Models
8/21/2014 30
32. a b c d abababababcdcdababcdcdcd
The number of true predictions = 0
a b c d
1.0 1.0 1.0
1.0
M1 M2 M3 M4
Hierarchical clustering
33. a b c d
ab
abababababcdcdababcdcdcd
The number of true predictions = 12
a b c d
1.0
1.0
1.0 1.0
M1
M2 M3
Hierarchical clustering
34. a b c d
ab
abababababcdcdababcdcdcd
The number of true predictions = 20
a b c d
1.0
1.0
1.0
1.0
cd
M1
M2
Hierarchical clustering
35. a b c d
ab
abababababcdcdababcdcdcd
The number of true predictions = 20
a b c d
1.0
1.0
1.0
1.0
cd
abcd
Stop as long as no additional prediction benefit of merging
M1
M2
Hierarchical clustering
37. Schema for Hieratical
Clustering
Web log
Train:
To train predictive
models
Validation:
To find Best
clusters
Test:
To derive final
accuracy
To find a “Best”
clusters
Calculate final
accuracy
To train local
predictive models
8/21/2014 37
38. Resulted Clusters
Id Summary Cluster
1 Intensive Search Basic Search, Refine Search, Empty
Search Result
2 Explore information
related to
program
Program impression in search result,
Banner click, Program click ,
Click on university link
3 Start of
browsing
University Spotlight impression,
Quick search
4 Explore
country information
File view, Click on country link
5 Explore
search result
Program impressions in search results,
University impression on
nearby universities
6 Explore program Program in landing page, Submit
inquiry
7 Outlier Submit question, X-node
8/21/2014 38
40. Site Map
• Page is represented as set of possible actions
o Example: Homepage is (Quick Search, University Spot light
impression, Question Submit)
o Calculate Jaccard similarity between Page and Cluster
8/21/2014 40
42. Conclusion
• We formulated the context discovery
process as an optimization problem
• Definition of the useful contexts
• Our approach can be used to identify
useful contexts for next action prediction
• Experiments on a real dataset provide
empirical evidence that our methods are
better than baselines
8/21/2014 42
43. Future works
• Testing our method on another datasets
• Introducing a mechanism for detecting
context switching within a web-session
• Considering multidimensional contextual
features
8/21/2014 43
44. Thank you!
• Context mining and integration into
prediction models
• Accurately predicting users’ desired
actions and understanding behavioral
patterns of users in various web-
applications
• Personalization and adaptation to
diverse customer needs and
preferences
• Accounting for the practical needs
within the considered application areas
8/21/2014 44
Many taxonomies were built for explictit context.
Physical Environment Factors -> Conditions -> Weather ->
Context can have different grannularity
Linking discovered context into predictive modeling
or
Context-aware adaptation of predictive modeling
Masterportal homepge to make you familiar with the portal
Screenshot
What is the next action prediction of the user?
The most common – make use of Markov Model.
The question what kind of context we can define and/or discover for this problem
In order to build a local Markov models we can partition our data.
There are two types of partitioning: horizontal and vertical. Let’s start from horizontal. As context we use “user geo location”.
Jaccard similarity between clusters and pages
Explore
Intensive search on search result page
Statement 1 line
----- Meeting Notes (7/28/14 15:14) -----
as an optimization
no reading - talk
----- Meeting Notes (7/28/14 15:14) -----
Other datasets
Another dataset
Alighment problem