Shadi Banitaan

Pareto efficient multi-objective optimization for local tuning of analogy-based estimation

Abstract Analogy-based effort estimation (ABE) is one of the prominent methods for software effort estimation. The fundamental concept of ABE is closer to the mentality of expert estimation but with an automated procedure in which the final estimate is generated by reusing similar historical projects. The main key issue when using ABE is how to adapt the effort of the retrieved nearest neighbors. The adaptation process is an essential part of ABE to generate more successful accurate estimation based on tuning the selected raw solutions, using some adaptation strategy. In this study, we show that there are three interrelated decision variables that have great impact on the success of adaptation method: (1) number of nearest analogies (k), (2) optimum feature set needed for adaptation and (3) adaptation weights. To find the right decision regarding these variables, one need to study all possible combinations and evaluate them individually to select the one that can improve all prediction evaluation measures. The existing evaluation measures usually behave differently, presenting sometimes opposite trends in evaluating prediction methods. This means that changing one decision variable could improve one evaluation measure while it is decreasing the others. Therefore, the main theme of this research is how to come up with best decision variables that improve adaptation strategy and thus the overall evaluation measures without degrading the others. The impact of these decisions together has not been investigated before; therefore, we propose to view the building of adaptation procedure as a multi-objective optimization problem. The Particle swarm optimization algorithm (PSO) is utilized to find the optimum solutions for such decision variables based on optimizing multiple evaluation measures. We evaluated the proposed approaches over 15 datasets and using four evaluation measures. After extensive experimentation, we found that: (1) predictive performance of ABE has noticeably been improved, (2) optimizing all decision variables together is more efficient than ignoring any one of them, and (3) optimizing decision variables for each project individually yields better accuracy than optimizing them for the whole dataset.

Bug Reports Prioritization: Which Features and Classifier to Use?

Large open source bug tracking systems receives large number of bug reports daily. Managing these huge numbers of incoming bug reports is a challenging task. Dealing with these reports manually consumes time and resources which leads to delaying the resolution of important bugs which are crucial and need to be identified and resolved earlier. Bug triaging is an important process in software maintenance. Some bugs are important and need to be fixed right away, whereas others are minor and their fixes could be postponed until resources are available. Most automatic bug assignment approaches do not take the priority of bug reports in their consideration. Assigning bug reports based on their priority may play an important role in enhancing the bug triaging process. In this paper, we present an approach to predict the priority of a reported bug using different machine learning algorithms namely Naive Bayes, Decision Trees, and Random Forest. We also investigate the effect of using two feature sets on the classification accuracy. We conduct experimental evaluation using open-source projects namely Eclipse and Fire fox. The experimental evaluation shows that the proposed approach is feasible in predicting the priority of bug reports. It also shows that feature-set-2 outperformsfeature-set-1. Moreover, both Random Forests and Decision Trees outperform Naive Bayes.

CLARM: an integrative approach for functional modules discovery

Functional module discovery aims to find well-connected subnetworks which can serve as candidate protein complexes. Advances in High-throughput proteomic technologies have enabled the collection of large amount of interaction data as well as gene expression data. We propose, CLARM, a clustering algorithm that integrates gene expression profiles and protein protein interaction network for biological modules discovery. The main premise is that by enriching the interaction network by adding interactions between genes which are highly co-expressed over a wide range of biological and environmental conditions, we can improve the quality of the discovered modules. Protein protein interactions, known protein complexes, and gene expression profiles for diverse environmental conditions from the yeast Saccharomyces cerevisiae were used for evaluate the biological significance of the reported modules. Our experiments show that the CLARM approach is competitive to well-established module discovery methods.

A formal study of classification techniques on entity discovery and their application to opinion mining

Entity discovery has become an important topic of study in recent years due to its wide range of applications. In this paper, we focus on examining the effectiveness of various classification techniques on entity discovery and their application to the opinion mining task. The initial and most important step in opinion mining is to identify and extract highly specific product related and opinion related entities from product reviews. We formulate this problem as a classification task and present a comprehensive study of classification techniques on identifying entities of interest. The impacts of linguistic features such as part-of-speech (POS), and context features such as surrounding contextual clues of words on the classification performance are carefully evaluated. The experimental results show that good classification performance is closely related to the use of classification techniques, linguistic features, and context features. The evaluation is presented based on processing the online product reviews from Amazon.

Towards Test Focus Selection for Integration Testing Using Method Level Software Metrics

The aim of integration testing is to uncover errors in the interactions between system modules. However, it is generally impossible to test all the interactions between modules because of time and cost constraints. Thus, it is important to focus the testing on the connections presumed to be more error-prone. The goal of this research is to guide quality assurance team wherein a software system to focus when they perform integration testing to save time and resources. In this work, we use method level metrics that capture both dependencies and internal complexity of methods. In addition, we build a tool that calculates the metrics automatically. We also propose an approach to select the test focus in integration testing. The main goal is to reduce the number of test cases needed while still detecting at least 80% of integration errors. We conducted an experimental study on several Java applications taken from different domains. Error seeding technique have been used for evaluation. The experimental results showed that our proposed approach is very effective for selecting the test focus in integration testing. It reduces considerably the number of required test cases while at the same time detects at least 80% of integration errors.

Feature-Based Test Focus Selection Technique Using Classes Connections Weight

Testing could cost more than fifty percent of all development cost, particularly integration testing consumes around eighty percent of testing cost. Integration testing aims to discover errors in the connections among classes which are collaborate and communicate in order to provide specific services. Though, testing all connections among classes is impractical because of the cost, effort and time constraints. Test focus selection might help testers to concentrate on the main and vital connections among classes which it could be the most error prone ones. The authors proposed approach amalgamates the static and dynamic analysis in order to detect, trace, and weight the connections among classes through method level communications. Their approach harnessed an open source tracing tool MUTT. The MUTT allows them to return all the methods in all classes that have been called respecting to any specific feature which has triggered by the system user. The experimental results reveal how the proposed approach achieves good mutation testing score on the systems under study.

Using Data Mining to Predict Possible Future Depression Cases

Received Oct 28, 2014 Revised Nov 14, 2014 Accepted Nov 26, 2014 Depression is a disorder characterized by misery and gloominess felt over a period of time. Some symptoms of depression overlap with other somatic illnesses implying considerable difficulty in diagnosing it. This paper contributes to its diagnosis through the application of data mining, namely classification, to predict patients who will most likely develop depression or are currently suffering from depression. Synthetic data was used for this study. To acquire the results, the popular suite of machine learning software, WEKA, was used. Keyword:

An Application of Classification and Class Decomposition to Use Case Point Estimation Method

Use Case Points (UCP) estimation method describes the process of computing the software project size and productivity from use case diagram elements. These metrics are then used to predict the project effort at early stage of software development. The main challenges with previous models are that they were constructed based on a very limited number of observations, and using limited productivity ratios. This paper presents a new approach to predict productivity from UCP environmental factors by applying classification with decomposition technique. A class decomposition provides a number of advantages to supervised learning algorithms through segmenting classes into more homogenous classes, and therefore, increase their diversity. The proposed model is constructed and validated over two datasets that have relatively sufficient number of observations. The accuracy results are promising and have potential to increase accuracy of early effort estimation.

DECOBA: Utilizing Developers Communities in Bug Assignment

Bug Tracking System (BTS) is public ally accessible which enables geographically distributed developers to follow the work of each other and contribute in bug fixing. Developer interactions through commenting on bug reports generate a developer social network that can be used to improve software development and maintenance activities. In large scale complex software projects, software maintenance requires larger groups to participate in its activities. Most previous bug assignments approaches assign only one developer to new bugs. However, bug fixing is a collaborative effort between several developers (i.e., many developers contribute their experience in fixing a bug report). In this work, we build developers social networks based on developers comments on bug reports and detect developers communities. We also assign a relevant community to each newly committed bug report. Moreover, we rank developers in each community based on their experience. An experimental evaluation is conducted on three open source projects namely Net Beans, Free desktop, and Mandriva. The results show that the detected communities are significantly connected with high density. They also show that the proposed approach achieves feasible accuracy of bug assignment.

Improving Functional Modules Discovery by Enriching Interaction Networks with Gene Profiles

We would like to thank the anonymous reviewers for their comments and suggestions. This publication was made possible by NIH grant number P20RR016471 from the INBRE program of the National Center for Research Resources.