Nitish Mittal

Women in computer science research: what is the bibliography data telling us?

The low participation by women authors in research is an important equity issue in Computer Science Research (CSR). There are various parameters and methodologies that can be used to measure the gender imbalance. In this paper, we present a study on gender gap, imbalance and women participation in CSR. We conduct our experiments on DBLP bibliographical database and analyze several years of publication dataset across various domains of CSR. We perform Exploratory Data Analysis on the bibliographical dataset and study the trend of gender imbalance over several years. We propose eight research questions across various facets and our results shows a significant gender imbalance in different sub-fields within CSR and low rate of women participation across various regions of world.

A glance at seven ACM SIGWEB series of conferences

ACM Special Interest Group on Hypertext and the Web (ACM SIGWEB) is a community of researchers and supporting organization of seven conferences primarily focusing on the World Wide Web, document engineering, digital libraries and linked media oriented researches. ACM SIGWEB group covers a variety of topics related to hypertext, hypermedia, social networks, information retrieval and web science. In this paper, we present a bibliometric analysis of the scientific publications and corresponding ACM metadata (also published as a research output) of seven conferences sponsored by ACM SIGWEB. We perform an exploratory data analysis on bibliography database and analyze several years of publication data (since the year of conference started until September 2015) of these conferences. We propose 5 research questions across various facets such as number of publications, authors and affiliation participation and conduct a statistical analysis of DBLP records and ACM metadata to answer those questions. Our results reveal that new SIGWEB conferences (started in or after year 2000) are growing much faster in terms of number of publications, authors and affiliation participation across various regions of world. We also find that HT, despite being the oldest SIGWEB conference the rate of number submissions received are relatively less in comparison to WSDM (new SIGWEB conference). Our analysis also reveals that CIKM is the most leading SIGWEB conference.

Got a Complaint?- Keep Calm and Tweet It!

Research shows that many public service agencies use Twitter to share information and reach out to the public. Recently, Twitter is also being used as a platform to collect complaints from citizens and resolve them in an efficient time and manner. However, due to the dynamic nature of the website and presence of free-form-text, manual identification of complaint posts is overwhelmingly impractical. We formulate the problem of complaint identification as an ensemble classification problem. We perform several text enrichment processes such as hashtag expansion, spell correction and slang conversion on raw tweets for identifying linguistic features. We implement a one-class SVM classification and evaluate the performance of various kernel functions for identifying complaint tweets. Our result shows that linear kernel SVM outperforms polynomial and RBF kernel functions and the proposed approach classifies the complaint tweets with an overall precision of \(76\,\%\). We boost the accuracy of our approach by performing an ensemble on all three kernels. Result shows that one-class parallel ensemble SVM classifier outperforms cascaded ensemble learning with a margin of approximately \(20\,\%\). By comparing the performance of each kernel against ensemble classifier, we provide an efficient method to classify complaint reports.

A scientometric analysis of 9 ACM SIGWEB cooperating conferences

ACM SIGWEB is one of the ACMs special interest groups (SIG) supporting communities and researches focused in the areas of hypermedia, world wide web, hypertext, digital libraries, document engineering and social networks. In adition to seven sponsored conferences, ACM SIG-WEB cooperates with various computer science conferences called as the SIGWEB Cooperating Conference. Compared to ACM SIGWEB sponsored conference, cooperating conferences keep undergoing addition and removal of conference more often based on the topics of interest and themes. Similar to SIGWEB sponsored conferences, cooperating conferences also publish in a wide range of topics in hypermedia and multi-disciplinary domains. In this paper, we present a bibliometric analysis of scientific publications records of 9 SIGWEB cooperating conferences. We use DBLP bibliographical database as a base database and enhance the DBLP records of cooperating conferences using AMiner and Google Scholar. We perform and exploratory and scientometric analysis on publications, authors and conference database of 10 years (2006-2015) of SIGWEB cooperating conferences. We propose 4 research questions around the topics of research community in cooperating conferences such as authors participation, number of publications and relative comparison of google scholar h-index with overall citations received on each article published in an year in cooperating conferences. Our results reveal that over the past decade conferences like ASONAM and IWCMC have an increment in growth rate of the conference every year. These conferences have been publishing a large number of articles with maximum number of authors participation from the community. Further, despite having a large number of articles, h5-index and overall citation rate, RecSys covers only a limited range of topics and focused in the domain of base topics in recommender systems only. For example, matrix factorization and collaborative filtering. On a contrary, our empirical analysis reveal that due to the focused topics and domain, articles published in RecSys has received maximum number of citations over the past decade. While, conferences like Compute and RuleML has very low citation rate and h5-index upto 14 which is reasonably lower than the h5-index of other conferences (greater than 25). We further study the research community of ACM SIGWEB cooperating conferences and find that despite having a small number of papers in ACM-SCopC community, many researchers make among top 20 prolific authors receiving maximum number of citations on their articles published in ACM-SCopC.

Integrating greedy based approach with genetic algorithm to generate mixed covering arrays for pair-wise testing

The effectiveness of combinatorial interaction testing (CIT) to test highly configurable systems has constantly motivated researchers to look out for new techniques to construct optimal covering arrays that correspond to test sets. Pair-wise testing is a combinatorial testing technique that generates a pair-wise interaction test set to test all possible combinations of each pair of input parameter value. Meta heuristic techniques have being explored by researchers in past to construct optimal covering arrays for t-way testing (where, t denotes the strength of interaction). In this paper we apply genetic algorithm, a meta heuristic search based optimization algorithm to generate optimal mixed covering arrays for pair-wise testing. Here, we present a novel method that uses a greedy based approach to perform mutation and study the impact of the proposed approach on the performance of genetic algorithm. We describe the implementation of the proposed approach by extending an open source tool PWiseGen. Experimental results indicate that the use of greedy approach to perform mutation improves the performance of genetic algorithm by generating mixed covering arrays with higher fitness level in less number of generations as compared to those generated using other techniques.

Potholes and bad road conditions: mining Twitter to extract information on killer roads

Research shows that Twitter is being used as a platform to not only share and disseminate the information but also collecting complaints from citizens. However, due to the presence of high volume and large stream data, real-time manual identification of those complaints is overwhelmingly impractical. In this paper, we identify the complaints and grievances posted on bad road conditions causing life risks, discomfort and poor road experience to the citizens. We formulate the problem of killer road complaints identification as a multiclass text classification problem. We address the challenge of keyword based flagging methods and identify several linguistic features that are unique for the killer road complaints tweets such as the issue reported in the complaint, pinpoint location of the issue, city or region location information. Our results reveal that not all complaint reports posted to Public agencies contain the sufficient information and are not useful. Therefore, we further propose a mechanism to enrich the nearly-useful tweets and convert them into useful reports. We present our results using information visualization and gain actionable insights from them. Our results show that the proposed features are discriminatory and able to classify killer roads complaints with an accuracy of 67% and a recall of 65%.

How healthy are ACM SIGWEB sponsored conferences

ACM SIGWEB sponsors and supports several conferences covering a wide range of topics. CIKM, HT, WEBSCI and WSDM are four of the many sponsored conferences by ACM SIGWEB. CIKM, HT, WEBSCI and WSDM are reputed conferences, being held for several years and has a large community of contributing authors and PC members. In this article, we present a study on the health of these four conferences based on several factors and metrics such as stability, openness, inbreeding, representativeness, sustainability, prestige and workload. Studying the health of a conferences provides a reflection and historical overview in-terms of its performance which can be used by the conference community to bring improvements and ensure that the conference is meeting its desired objectives. We conduct statistical analysis and information visualization on the conference data downloaded from DBLP, ACM Digital Library web-pages and conference websites. Our analysis reveals that overall all the four conferences are showing good health indicators. We do observe variances in metrics values across conferences and within conferences across years.

Construction of Variable Strength Covering Array for Combinatorial Testing Using a Greedy Approach to Genetic Algorithm

The limitation of time and budget usually prohibits exhaustive testing of interactions between components in a component based software system. Combinatorial testing is a software testing technique that can be used to detect faults in a component based software system caused by the interactions of components in an eective and ecient way. Most of the research in the field of combinatorial testing till now has focused on the construction of optimal covering array (CA) of fixed strength t which covers all t-way interactions among components. The size of CA increases with the increase in strength of testing t, which further increases the cost of testing. However, not all components require higher strength interaction testing. Hence, in a system with k components a technique is required to construct CA of fixed strength t which covers all t-way interactions among k components and all ti-way (where ti > t) interactions between a subset of k components. This is achieved using the variable strength covering array (VSCA). In this paper we propose a greedy based genetic algorithm (GA) to generate optimal VSCA. Experiments are conducted on several benchmark configurations to evaluate the eectiveness of the proposed approach.

Analysis of a Sporting Event on a Social Network: True Popularity & Popularity Bond

Events in a social network and their popularity are described by the quantitative participation of its users. A special occasion in an event is an activity that may hamper or strengthen its popularity or popularity of its entities. Such a study helps the researchers to analyze the trends to know how they change with time during an occasion. Till now popularity is computed by considering the number of tweets. To the best of our knowledge, no study has been done on computing the number of tweets considering the population. Here in this paper, we coin the following terms (a) true popularity, which is the number of tweets normalized with the population. Through this we compare intra-group popularity of entities, and (b) popularity bond, so as to study concentration of tweets for pairs of entities. Through this we compare the inter-group popularity of entities. Experiments are carried out on the content posted by users on Twitter during the Cricket World Cup 2015. Experimental study indicates the effectiveness of the coined terms in providing better insights.

A general overview and bibliometric analysis of seven ACM hypertext and web conferences

Bibliometric analysis of published scientific papers is a widely used practice to conduct quantitative evaluations and assessments of conferences. In this study, we performed an in-depth bibliometric, scientometric and exploratory analysis of ACM SIGWEB sponsored conferences by visually analysing the DBLP database. We conducted a series of experiments and empirical analysis to answer several questions. Our results showed that the articles published in SIGWEB conferences stem from a variety of countries while the degree of cross-country collaboration is relatively low and that most co-authors of publications are by researchers who all reside in the same country. Collectively, SIGWEB conferences have a higher hosting rate and local community participation in places where the USA and Europe make the greatest conference contributions. Our results showed that the participation of female authors in SIGWEB conferences is increasing while in contrast, there is a huge gender imbalance in leadership and official conference positions.