Ashesh Mahidadia

Collaborative filtering for people to people recommendation in social networks

Predicting people other people may like has recently become an important task in many online social networks. Traditional collaborative filtering approaches are popular in recommender systems to effectively predict user preferences for items. However, in online social networks people have a dual role as both “users” and “items”, e.g., both initiating and receiving contacts. Here the assumption of active users and passive items in traditional collaborative filtering is inapplicable. In this paper we propose a model that fully captures the bilateral role of user interactions within a social network and formulate collaborative filtering methods to enable people to people recommendation. In this model users can be similar to other users in two ways – either having similar “taste” for the users they contact, or having similar “attractiveness” for the users who contact them. We develop SocialCollab, a novel neighbour-based collaborative filtering algorithm to predict, for a given user, other users they may like to contact, based on user similarity in terms of both attractiveness and taste. In social networks this goes beyond traditional, merely taste-based, collaborative filtering for item selection. Evaluation of the proposed recommender system on datasets from a commercial online social network show improvements over traditional collaborative filtering.

Interaction-based collaborative filtering methods for recommendation in online dating

We consider the problem of developing a recommender system for suggesting suitable matches in an online dating web site. The main problem to be solved is that matches must be highly personalized. Moreover, in contrast to typical product recommender systems, it is unhelpful to recommend popular items: matches must be extremely specific to the tastes and interests of the user, but it is difficult to generate such matches because of the two way nature of the interactions (user initiated contacts may be rejected by the recipient). In this paper, we show that collaborative filtering based on interactions between users is a viable approach in this domain. We propose a number of new methods and metrics to measure and predict potential improvement in user interaction success, which may lead to increased user satisfaction with the dating site. We use these metrics to rigorously evaluate the proposed methods on historical data collected from a commercial online dating web site. The evaluation showed that, had users been able to follow the top 20 recommendations of our best method, their success rate would have improved by a factor of around 2.3.

Reciprocal and heterogeneous link prediction in social networks

Link prediction is a key technique in many applications in social networks, where potential links between entities need to be predicted. Conventional link prediction techniques deal with either homogeneous entities, e.g., people to people, item to item links, or non-reciprocal relationships, e.g., people to item links. However, a challenging problem in link prediction is that of heterogeneous and reciprocal link prediction, such as accurate prediction of matches on an online dating site, jobs or workers on employment websites, where the links are reciprocally determined by both entities that heterogeneously belong to disjoint groups. The nature and causes of interactions in these domains makes heterogeneous and reciprocal link prediction significantly different from the conventional version of the problem. In this work, we address these issues by proposing a novel learnable framework called ReHeLP , which learns heterogeneous and reciprocal knowledge from collaborative information and demonstrate its impact on link prediction. Evaluation on a large commercial online dating dataset shows the success of the proposed method and its promise for link prediction.

Contents and time sensitive document ranking of scientific literature

A new link-based document ranking framework is devised with at its heart, a contents and time sensitive random literature explorer designed to more accurately model the behaviour of readers of scientific documents. In particular, our ranking framework dynamically adjusts its random walk parameters according to both contents and age of encountered documents, thus incorporating the diversity of topics and how they evolve over time into the score of a scientific publication. Our random walk framework results in a ranking of scientific documents which is shown to be more effective in facilitating literature exploration than PageRank measured against a proxy gold standard based on papers’ potential usefulness in facilitating later research. One of its many strengths lies in its practical value in reliably retrieving and placing promisingly useful papers at the top of its ranking.

Assisting Model-Discovery in Neuroendocrinology

It is very difficult, if not impossible, for researchers to manually evaluate and revise their scientific models using a vast amount of relevant information now available to them. The paper describes a new framework, called JustAid, that successfully integrates techniques from Knowledge Acquisition and Machine Learning in a way that complements their strengths to overcome their weaknesses, and provides an interactive environment to help researchers in a process of scientific discovery. JustAid can use information stored in medical databases and assist experimental scientists in forming, testing and revising scientific models, without a need of a knowledge engineer. In this paper, JustAid has been applied to a real world problem in the area of neuroendocrinology, a branch of physiology.

Feature extraction for learning to classify questions

In this paper, we present a new approach to learning the classification of questions Question classification received interest recently in the context of question answering systems for which categorizing a given question would be beneficial to allow improved processing of the document to identify an answer Our approach relies on relative simple preprocessing of the question and uses standard decision tree learning We also compared our results from decision tree learning with those obtained using Naive Bayes Both results compare favorably to several very recent studies using more sophisticated preprocessing and/or more sophisticated learning techniques Furthermore, the fact that decision tree learning proved more successful than Naive Bayes is significant in itself as decision tree learning is usually believed to be less suitable for NLP tasks.

People-to-People recommendation using multiple compatible subgroups

People-to-people recommendation aims at suggesting suitable matches to people in a way that increases the likelihood of a positive interaction. This problem is more difficult than conventional item-to-people recommendation since the preferences of both parties need to be taken into account. Previously we proposed a profile-based recommendation method that first uses compatible subgroup rules to select a single best attribute value for each corresponding value of the user, then combines these attribute value pairs into a rule that determines the recommendations. Though this method produces a significant improvement in the probability of an interaction being successful, it has two significant limitations: (i) by considering only single matching attribute values the method ignores cases where different attribute values are closely related, missing potential candidates, and (ii) when ranking candidates for recommendation the method does not consider individual behaviour. This paper addresses these two issues, showing how multiple attributes can be used with compatible subgroup rules and individual reply rates used for ranking candidates. Our experimental results show that the new approach significantly improves the probability of an interaction being successful compared to our previous approach.

Mining patterns of dyspepsia symptoms across time points using constraint association rules

In this paper, we develop and implement a framework for constraint-based association rule mining across subgroups in order to help a domain expert find useful patterns in a medical data set that includes temporal data. This work is motivated by the difficulties experienced in the medical domain to identify and track dyspepsia symptom clusters within and across time. Our framework, Apriori with Subgroup and Constraint (ASC), is built on top of the existing Apriori framework. We have identified four different types of phase-wise constraints for subgroups: constraint across subgroups, constraint on subgroup, constraint on pattern content and constraint on rule. ASC has been evaluated in a real-world medical scenario; analysis was conducted with the interaction of a domain expert. Although the framework is evaluated using a data set from the medical domain, it should be general enough to be applicable in other domains.

Learning to make social recommendations: a model-based approach

Social recommendation, predicting people who match other people for friendship or as potential partners in life or work, has recently become an important task in many social networking sites. Traditional content-based and collaborative filtering methods are not sufficient for people-to-people recommendation because a good match depends on the preferences of both sides. We proposed a framework for social recommendation and develop a representation for classification of interactions in online dating applications that combines content from user profiles plus interaction behaviours. We show that a standard algorithm can be used to learn a model to predict successful interactions. We also use a method to search for the best model by minimising a cost based on predicted precision and recall. To use the model in real world applications to make recommendations, we generate candidate pairs using the selected models and ranked them using a novel probabilistic ranking function to score the chance of success. Our model-based social recommender system is evaluated on historical data from a large commercial social networking site and shows improvements in success rates over both interactions with no recommendations and those with recommendations generated by standard collaborative filtering.

Knowledge Management in Data and Knowledge Intensive Environments

In this digital age, it is now possible to electronically collect a large amount of data and business knowledge. However, the next crucial step is to make sense of these data and knowledge in order to improve business processes. Unfortunately, most of the tools available today are not capable of evaluating a large amount of available data against the current business knowledge, in order to automatically suggest improvements and help a decision maker in the process of revising current business processes. In this paper, we outline a new framework that assists a decision maker in the process of evaluating and then if required revising current business knowledge. The tool presented in this paper has been successfully applied to test and revise knowledge bases in the medical domain, using real world data and a domain expert.