Bashar Al-Shboul

A Genetic Programming Based Framework for Churn Prediction in Telecommunication Industry

Customer defection is critically important since it leads to serious business loss. Therefore, investigating methods to identify defecting customers (i.e. churners) has become a priority for telecommunication operators. In this paper, a churn prediction framework is proposed aiming at enhancing the ability to forecast customer churn. The framework combine two heuristic approaches: Self Organizing Maps (SOM) and Genetic Programming (GP). At first, SOM is used to cluster the customers in the dataset, and then remove outliers representing abnormal customer behaviors. After that, GP is used to build an enhanced classification tree. The dataset used for this study contains anonymized real customer information provided by a major local telecom operator in Jordan. Our work shows that using the proposed method surpasses various state-of-the-art classification methods for this particular dataset.

Wikipedia-based query phrase expansion in patent class search

Relevance feedback methods generally suffer from topic drift caused by word ambiguities and synonymous uses of words. Topic drift is an important issue in patent information retrieval as people tend to use different expressions describing similar concepts causing low precision and recall at the same time. Furthermore, failing to retrieve relevant patents to an application during the examination process may cause legal problems caused by granting an existing invention. A possible cause of topic drift is utilizing a relevance feedback-based search method. As a way to alleviate the inherent problem, we propose a novel query phrase expansion approach utilizing semantic annotations in Wikipedia pages, trying to enrich queries with phrases disambiguating the original query words. The idea was implemented for patent search where patents are classified into a hierarchy of categories, and the analyses of the experimental results showed not only the positive roles of phrases and words in retrieving additional relevant documents through query expansion but also their contributions to alleviating the query drift problem. More specifically, our query expansion method was compared against relevance-based language model, a state-of-the-art query expansion method, to show its superiority in terms of MAP on all levels of the classification hierarchy.

Analyzing topic drift in query expansion for Information Retrieval from a large-scale patent DataBase

Topic drift has been recognized as a major reason for ineffective retrieval of documents using query expansion. Topic drift is very important in patent domain as document processing is dependable on the quality of the search. With the huge amount of available patents, we investigate on the concept of topic drift in patent domain by analyzing the topics of retrieval results obtainable with query expansion. We define and utilize topic attributes to distinguish drifting topics from other topic classes such as Vanishing, Appearing, and Rising Topics. As a result, this analysis can be used to compare two different query expansion methods for their relative effectiveness.

Exploiting Social Media and Tagging for Social Book Search: Simple Query Methods for Retrieval Optimization

With web and social media information availability and accessibility increasing on the one hand, there is an increased complexity of retrieval on the other. Generally, real information needs are complex to express. Books are the prevailing information resources and book searching is one of the online activities that users attempt daily. Mobile technology makes it easy to handle books, i.e., ibook and kindle formats; however, locating users’ preference over the Internet is still quite complex. Efforts are being made to help users locate their desired books easily and quickly. This research is set up to investigate techniques to support users in searching and navigating the full texts of digitized books and complementary social media in order to enhance the user book search experience. The idea is based on the INEX SBS track to use professional metadata and user-generated metadata to enhance the retrieval process of books by optimizing simple search query with INEX SBS 2015. Amazon and LibraryThing book descriptions were processed to extract information and important fields to be indexed. The proposed model use the Named Entity Recognition tagger (NER) and the Part-Of-Speech tagger (POS) to extract relevant topics that are related to book search. The results indicated that using simple methods such as NER and POS tagging can generate an effective query for book retrieval.

A Hybrid Approach Based on Particle Swarm Optimization and Random Forests for E-Mail Spam Filtering

Internet is flooded every day with a huge number of spam emails. This will lead the internet users to spend a lot of time and effort to manage their mailboxes to distinguish between legitimate and spam emails, which can considerably reduce their productivity. Therefore, in the last decade, many researchers and practitioners proposed different approaches in order to increase the effectiveness and safety of spam filtering models. In this paper, we propose a spam filtering approach consisted of two main stages; feature selection and emails classification. In the first step a Particle Swarm Optimization (PSO) based Wrapper Feature Selection is used to select the best representative set of features to reduce the large number of measured features. In the second stage, a Random Forest spam filtering model is developed using the selected features in the first stage. Experimental results on real-world spam data set show the better performance of the proposed method over other five traditional machine learning approaches from the literature. Furthermore, four cost functions are used to evaluate the proposed spam filtering method. The results reveal that the PSO based Wrapper with Random Forest can effectively be used for spam detection.