Zeeshan Ali Rana

Ineffectiveness of Use of Software Science Metrics as Predictors of Defects in Object Oriented Software

Software science metrics (SSM) have been widely used as predictors of software defects. The usage of SSM is an effect of correlation of size and complexity metrics with number of defects. The SSM have been proposed keeping in view the procedural paradigm and structural nature of the programs. There has been a shift in software development paradigm from procedural to object oriented (OO) and SSM have been used as defect predictors of OO software as well. However, the effectiveness of SSM in OO software needs to be established. This paper investigates the effectiveness of use of SSM for: a)classification of defect prone modules in OO software b) prediction of number of defects. Various binary and numeric classification models have been applied on dataset kc1 with class level data to study the role of SSM. The results show that the removal of SSM from the set of independent variables does not significantly affect the classification of modules as defect prone and the prediction of number of defects. In most of the cases the accuracy and mean absolute error has improved when SSM were removed from the set of independent variables. The results thus highlight the ineffectiveness of use of SSM in defect prediction in OO software.

Finding focused itemsets from software defect data

Software product measures have been widely used to predict software defects. Though these measures help develop good classification models, studies propose that relationship between software measures and defects still needs to be investigated. This paper investigates the relationship between software measures and the defect prone modules by studying associations between the two. The paper identifies the critical ranges of the software measures that are strongly associated with defects across five datasets of PROMISE repository. The paper also identifies the ranges of the measures that do not necessarily contribute towards defects. These results are supported by information gain based ranking of software measures.

Towards a generic model for software quality prediction

Various models and techniques have been proposed and applied in literature for software quality prediction. Specificity of each suggested model is one of the impediments in development of a generic model. A few models have been quality factor specific whereas others are software development paradigm specific. The models can even be company specific or domain specific. The amount of work done for software quality prediction compels the researchers to get benefit from the existing models and develop a relatively generic model. Development of a generic model will facilitate the quality managers by letting them focus on how to improve the quality instead of employing time on deciding which technique best suites their scenario. This paper suggests a generic model which takes software as input and predicts a quality factor value using existing models. This approach captures the specificity of existing models in various dimensions (like quality factor, software development paradigm, and software development life cycle phase etc.), and calculates quality factor value based on the model with higher accuracy. Application of the model has been discussed with the help of an example.

Improving Recall of software defect prediction models using association mining

Use of software product metrics in defect prediction studies highlights the utility of these metrics. Public availability of software defect data based on the product metrics has resulted in the development of defect prediction models. These models experience a limitation in learning Defect-prone (D) modules because the available datasets are imbalanced. Most of the datasets are dominated by Not Defect-prone (ND) modules as compared to D modules. This affects the ability of classification models to learn the D modules more accurately. This paper presents an association mining based approach that allows the defect prediction models to learn D modules in imbalanced datasets. The proposed algorithm preprocesses data by setting specific metric values as missing and improves the prediction of D modules. The proposed algorithm has been evaluated using 5 public datasets. A Naive Bayes (NB) classifier has been developed before and after the proposed preprocessing. It has been shown that Recall of the classifier after the proposed preprocessing has improved. Stability of the approach has been tested by experimenting the algorithm with different number of bins. The results show that the algorithm has resulted in up to 40% performance gain.

Identifying Association between Longer Itemsets and Software Defects

Software defects are an indicator of software quality. Software with lesser number of defective modules are desired. Prediction of software defects using software measurements facilitates early identification of defect-prone modules. Association relationship between software measures and defects improves prediction of defective modules. To find association relationship between software measures and defects, each numeric measure is divided into bins. Each bin is called 1-itemset (or an itemset of length 1). When certain itemsets and defective modules appear together in a dataset, they are considered associated with each other. Frequency of their co-occurrence depicts the strength of the association relationship. Existing studies find the relationship between 1-itemsets and defective modules. Itemsets that have high association with defects are called focused itemsets. Focused itemsets can be used to build prediction models with higher Recall values. This paper explores the relationship between defective modules and itemsets with length greater than 1. Focused itemsets with length greater than 1 involve multiple bins at same time. Identification of the focused itemsets has improved the performance of decision tree based defect prediction model.

Nomenclature unification of software product measures

A large number of software quality prediction models are based on software product measures (SPdMs). There are different interpretations and representations of these measures which generate inconsistencies in their naming conventions. These inconsistencies affect the efforts to develop a generic approach to predict software quality. This study identifies two types of such inconsistencies and categorises them into Type I and Type II. Type I inconsistency emerges when different labels are suggested for the same software product measure. Type II inconsistency appears when same label is used for different measures. This study suggests a unification and categorisation framework to remove Type I and Type II inconsistencies. The proposed framework categorises SPdMs with respect to three dimensions: usage frequency, software development paradigm and software lifecycle phase. The framework is applied to 140 SPdMs and a searchable unified measures database (UMD) is developed. Overall, 48.5% of the measures are found inconsistent. Out of the total measures studied 34.28% measures are frequently used. It has been found that 30.71% measures are used in object oriented paradigm and 31.43% measures are used in conventional paradigm. There is an overlap of 37.86% measures between the two paradigms. The UMD reveals that the percentages of measures used in design and implementation phases are 52.86 and 35%, respectively.

An FIS for early detection of defect prone modules

Early prediction of defect prone modules helps in better resource planning, test planning and reducing the cost of defect correction in later stages of software lifecycle. Early prediction models based on design and code metrics are difficult to develop because precise values of the model inputs are not available. Conventional prediction techniques require exact inputs, therefore such models cannot always be used for early predictions. Innovative prediction methods that use imprecise inputs, however, can be applied to overcome the requirement of exact inputs. This paper presents a fuzzy inference system (FIS) that predicts defect proneness in software using vague inputs defined as fuzzy linguistic variables. The paper outlines the methodology for developing the FIS and applies the model to a real dataset. Performance analysis in terms of recall, accuracy, misclassification rate and a few other measures has been conducted resulting in useful insight to the FIS application. The FIS model predictions at an early stage have been compared with conventional prediction methods (i.e. classification trees, linear regression and neural networks) based on exact values. In case of the FIS model, the maximum and the minimum performance shortfalls were noticed for true negative rate (TNRate) and F measure respectively. Whereas for Recall, the FIS model performed better than the other models even with the imprecise inputs.

Using Association Rules to Identify Similarities between Software Datasets

A number of V&V datasets are publicly available. These datasets have software measurements and defectiveness information regarding the software modules. To facilitate V&V, numerous defect prediction studies have used these datasets and have detected defective modules effectively. Software developers and managers can benefit from the existing studies to avoid analogous defects and mistakes if they are able to find similarity between their software and the software represented by the public datasets. This paper identifies the similar datasets by comparing association patterns in the datasets. The proposed approach finds association rules from each dataset and identifies the overlapping rules from the 100 strongest rules from each of the two datasets being compared. Afterwards, average support and average confidence of the overlap is calculated to determine the strength of the similarity between the datasets. This study compares eight public datasets and results show that KC2 and PC2 have the highest similarity 83% with 97% support and 100% confidence. Datasets with similar attributes and almost same number of attributes have shown higher similarity than the other datasets.

A comparative study of spatial complexity metrics and their impact on maintenance effort

Maintenance is an important phase of software lifecycle. A significant fraction of overall effort and cost is consumed on comprehension and maintenance. Maintenance effort is highly correlated with spatial complexity of the system. The existing metrics for measuring spatial complexity are not suitable for all types of software systems. These metrics sometimes do not depict the effect of the key factors, which contribute significantly towards maintenance effort. In addition, most of these metrics are suitable for a specific type of a system. In this paper, the existing metrics have been studied and their application areas have been identified

EnerPlan: smart energy management planning for home users

The impending energy crisis has driven up the cost of electricity at an exponential rate. Managing electric consumption thus has become a very crucial task especially for home consumers. In this paper we present EnerPlan, a non-intrusive method to aid consumers to reduce their energy cost by advising them a consumption plan for their devices. Our system builds consumer classes based on regional statistical data. Using these classes a target consumers device load and distribution is inferred. This inferred data is used to construct a device usage plan. Following this plan can reduce the electric bill of the consumer. We use expert-based and auto-generated fuzzy rules to generate the planning. Results show that in absence of experts, planning through auto-generated is also useful. The results further demonstrate that the data prepared using the proposed approach can be used to save electricity and the plans generated by EnerPlan can reduce electricity bills of consumers.