Touhid Bhuiyan

Optimal Trust Network Analysis with Subjective Logic

Trust network analysis with subjective logic (TNA-SL) simplifies complex trust graphs into series-parallel graphs by removing the most uncertain paths to obtain a canonical graph. This simplification could in theory cause loss of information and thereby lead to sub-optimal results. This paper describes a new method for trust network analysis which is considered optimal because it does not require trust graph simplification, but instead uses edge splitting to obtain a canonical graph. The new method is compared with TNA-SL, and our simulation shows that both methods produce equal results. This indicates that TNA-SL in fact also represents an optimal method for trust network analysis and that the trust graph simplification does not affect the result.

Combining Trust and Reputation Management for Web-Based Services

Services offered and provided through the Web have varying quality, and it is often difficult to assess the quality of a services before accessing and using it. Trust and reputation systems can be used in order to assist users in predicting and selecting the best quality services. This paper describes how Bayesian reputation systems can be combined with trust modeling based on subjective logic to provide an integrated method for assessing the quality of online services. This will not only assist the users decision making, but will also provide an incentive for service providers to maintain high quality, and can be used as a sanctioning mechanism to discourage deceptive and low quality services.

An Analysis of Trust Transitivity Taking Base Rate into Account

Trust transitivity, as trust itself, is a human mental phenomenon, so there is no such thing as objective transitivity, and trust transitivity therefore lends itself to different interpretations. Trust transitivity and trust fusion both are important elements in computational trust. This paper analyses the parameter dependence problem in trust transitivity and proposes some definitions considering the effects of base rate. In addition, it also proposes belief functions based on subjective logic to analyze trust transitivity of three specified cases with sensitive and insensitive based rate. Then it presents a quantitative analysis of the issue of exaggerated beliefs in Mass Hysteria based on subjective logic.

Developing Trust Networks Based on User Tagging Information for Recommendation Making

Recommender systems are one of the recent inventions to deal with ever growing information overload. Collaborative filtering seems to be the most popular technique in recommender systems. With sufficient background information of item ratings, its performance is promising enough. But research shows that it performs very poor in a cold start situation where previous rating data is sparse. As an alternative, trust can be used for neighbor formation to generate automated recommendation. User assigned explicit trust rating such as how much they trust each other is used for this purpose. However, reliable explicit trust data is not always available. In this paper we propose a new method of developing trust networks based on users interest similarity in the absence of explicit trust data. To identify the interest similarity, we have used users personalized tagging information. This trust network can be used to find the neighbors to make automated recommendations. Our experiment result shows that the proposed trust based method outperforms the traditional collaborative filtering approach which uses users rating data. Its performance improves even further when we utilize trust propagation techniques to broaden the range of neighborhood.

Trust for Intelligent Recommendation

Recommender systems are one of the recent inventions to deal with the ever-growing information overload in relation to the selection of goods and services in a global economy. Collaborative Filtering (CF) is one of the most popular techniques in recommender systems. The CF recommends items to a target user based on the preferences of a set of similar users known as the neighbors, generated from a database made up of the preferences of past users. In the absence of these ratings, trust between the users could be used to choose the neighbor for recommendation making. Better recommendations can be achieved using an inferred trust network which mimics the real world friend of a friend recommendations. To extend the boundaries of the neighbor, an effective trust inference technique is required. This book proposes a trust interference technique called Directed Series Parallel Graph (DSPG) that has empirically outperformed other popular trust inference algorithms, such as TidalTrust and MoleTrust. For times when reliable explicit trust data is not available, this book outlines a new method called SimTrust for developing trust networks based on a users interest similarity. To identify the interest similarity, a users personalized tagging information is used. However, particular emphasis is given in what resources the user chooses to tag, rather than the text of the tag applied. The commonalities of the resources being tagged by the users can be used to form the neighbors used in the automated recommender system. Through a series of case studies and empirical results, this book highlights the effectiveness of this tag-similarity based method over the traditional collaborative filtering approach, which typically uses rating data. Trust for Intelligent Recommendation is intended for practitioners as a reference guide for developing improved, trust-based recommender systems. Researchers in a related field will also find this book valuable.

SQLi penetration testing of financial Web applications: Investigation of Bangladesh region

Business critical web applications are the most popular services provided to the client by the financial sector. These applications are bringing handsome revenue for the financial industry every year. These services are also a frequent target of attackers. Poor coding practice leads applications to vulnerability that are exploited by attackers. Information and privileges such as access to databases, admin authorization, and access to data could be retrieved through exploitation. Services provided through web applications make the exploitation easier as these could be accessed from anywhere around the world. Web based financial services are comparatively new concept in Bangladesh. Thus the security aspects of these applications are less explored. This paper represents an analysis of few basic security issues of the financial web applications of Bangladesh. It focuses on structured query language injection (SQLi) vulnerability. It presents a manual black box penetration testing approach to test the financial web applications. Same steps are used for testing all the web applications in the dataset. A vulnerability analysis of the findings collected during the penetration testing is also presented in the paper.

Association Assessment among Risk Factors and Breast Cancer in a Low Income Country: Bangladesh

BACKGROUND In the low incoming country Bangladesh, breast cancer is second most common neoplasm and is increasing at an alarming rate among females. Lack of awareness and illiteracy are contributory factors for late presentation and therefore mortality. PURPOSE To examine associations of different factors with breast cancer mortality and to raise awareness among the women of society in Bangladesh. MATERIALS AND METHODS This descriptive case-control study was conducted on 160 participants from April 2011 till July 2014. Through a valid questionnaire covering personal and family history, data were collected by face to face interview. For analyzing correlations among factors with breast cancer data, binary logistic regression, Pearsons χ2- value, odd ratios and p-value tests were conducted with SPSS version 20. RESULTS The mean age of the patients was 43.0 (SD= ± 11.12). In ascending order the leading significant factors were hormone therapy (p<0.0000, OR=4.897), abortion (p<0.0001, OR=3.452), early start menarche (p<0.0002, OR=3.500), family history (p<0.0022, OR=3.235), and late menopause (p<0.0093, OR=3.674) with both χ2 test and logistic regression analyses. Non-significant factors were cancer experience, fatty food habits, marital status and taking alcohol. CONCLUSIONS Regarding the investigation of this study, significant and insignificant factors correlation visualization with breast cancer will be helpful to increase awareness among Bangladeshi women as well as all over the world.

SQLi vulnerabilty in education sector websites of Bangladesh

Bangladesh has announced every Government & Non -Government school and colleges must website. The Web sites have to include all data and information every school and colleges. The goal of this initiative is to ensure equal quality of education and to provide education to the remote areas of the country. Though is a very new concept yet an appreciable number of institutes have already started shifting their systems online. While this advancement is commendable yet there are drawbacks such as security risks of these Web sites and the data in them. One of the easiest yet treacherous security risks of website is SQLi. This paper focuses on various types of SQLi vulnerabilities such as: normal, error based double query, and blind injection techniques and their aggression on the educational Web sites of Bangladesh. Manual penetration testing with black box approach has been implemented in number of Web applications to check the vulnerabilities. The data found has been analyzed to draw statistical conclusion of the present condition of the educational Web sites of Bangladesh.

Evaluation of the Effectiveness of a Web-Based e-Learning System for Tertiary Educational Institution

Thegrowing demand of e-learning system in academia is well accepted worldwide. One of the vital missions of e-learning is to allow people to learn for personal accomplishment or to earn a professional degree, without physically attending a traditional university or academic setting. E-learning can be applied for all levels of schooling from grade school to graduate degrees, and is versatile enough to accommodate all learning styles. A good number of researchers all over the world have already worked for developing different types of web based e-learning system. In this paper, we have presented our proposedweb-based e-learning system which is essential for any tertiary educational institution. Both students and teachers are tied up in this system. Overall its a guided learning for student and a great way to change our typical learning system but not replacement of teaching, its only the supplement idea for teaching which helps to learn easily and improve the productivity. We have also evaluated the effectiveness of our proposed system by conducting a survey, analyzed and presented the result.

Managing and Analysing Software Product Line Requirements

Modelling software product line (SPL) features plays a crucial role to a successful development of SPL. Feature diagram is one of the widely used notations to mo del SPL variants. However, there is a lack of precisely defined formal notations for representing and verifying such models. This paper presents an approach that we adopt to model SPL variants by using UML and subsequently verify them by using first order logic. UML provides an overall mode lling view of the system. First-order logic provides a precise and rigorous interpretation of the feature diagrams. We model variants and their dependencies by using propositional connectives and build logicalexpressions. These expressions are then validated by the Alloy verification tool. The analysis and verification process is illustrated by using Computer Aided Dispatch (CAD) system.