An empirical study for method level refactoring prediction by ensemble technique and SMOTE to improve its efficiency

Code refactoring is the modification of structure with out altering its functionality. The refactoring task is critical for enhancing the qualities for non-functional attributes, such as efficiency, understandability, reusability, and flexibility. Our research aims to build an optimized model for refactoring prediction at the method level with 7 ensemble techniques and verities of SMOTE techniques. This research has considered 5 open source java projects to investigate the accuracy of our anticipated model, which forecasts refactoring applicants by the use of ensemble techniques (BAG-KNN, BAG-DT, BAG-LOGR, ADABST, EXTC, RANF, GRDBST). Data imbalance issues are handled using 3 sampling techniques (SMOTE, BLSMOTE, SVSMOTE) to improve refactoring prediction efficiency and also focused all features and significant features. The mean accuracy of the classifiers like BAG- DT is 99.53% ,RANF is 99.55%, and EXTC is 99.59. The mean accuracy of the BLSMOTE is 97.21%. The performance of classifiers and sampling techniques are shown in terms of the box-plot diagram.

DynComm

The analysis of dynamics in networks represents a great deal in the Social Network Analysis research area. To support students, teachers, developers, and researchers in this work, we introduce a novel R package, namely DynComm. It is designed to be a multi-language package used for community detection and analysis on dynamic networks. The package introduces interfaces to facilitate further developments and the addition of new and future developed algorithms to deal with community detection in evolving networks. This new package aims to abstract the programmatic interface of the algorithms, whether they are written in R or other languages, and expose them as functions in R.

Comprehensive Method of Botnet Detection Using Machine Learning

The botnet interrupts network devices and keeps control of the connections with the command, which controls the programmer, and the programmer controls the malicious code injected in the machine for obtaining information about the machines. The attacker uses a botnet to commence dangerous attacks as DDoS, phishing, despoil of information, and spamming. The botnet establishes with a large network and several hosts belong to it. In the paper, the authors proposed the framework of botnet detection by using an Artificial Neural Network. The author research upgrading the extant system by comprising of cache memory to fast the process. Finally, for detection, the author used an analytical approach, which is known as an artificial neural network that contains three layers: the input layer, hidden layer, output layer, and all layers are connected to correlate and approximate the results. The experiment result determines that the classifier with 25 epochs gives optimal accuracy is 99.78 percent and shows the detection rate is 99.7 percent.

A Software Fault Prediction on Inter and Intra Release Prediction Scenarios

Software quality engineering applied numerous techniques for assuring the quality of software, namely testing, verification, validation, fault tolerance, and fault prediction of the software. The machine learning techniques facilitate the identification of software modules as faulty or non-faulty. In most of the research, these approaches predict the fault-prone module in the same release of the software. Although, the model is found to be more efficient and validated when training and tested data are taken from previous and subsequent releases of the software respectively. The contribution of this paper is to predict the faults in two scenarios i.e. inter and intra release prediction. The comparison of both intra and inter-release fault prediction by computing various performance matrices using machine learning methods shows that intra-release prediction is having better accuracy compared to inter-releases prediction across all the releases. Also, but both the scenarios achieve good results in comparison to existing research work.

Towards a Formal Approach for Assessing the Design Quality of Object-Oriented Systems

The cost of software maintenance is always increasing. The companies are often confronted to failures and software errors. The quality of software to use is so required. In this paper, the authors propose a new formal approach for assessing the quality of object-oriented system design according to the quality assessment model. This approach consists in modeling the input software system by an automaton based on object-oriented design metrics and their relationship with the quality attributes. The model exhibits the importance of metrics through their links with the attributes of software quality. In addition, it is very practical and flexible for all changes. It allows the quality estimation and its validation. For the verification of proposed probabilistic model (automaton), they use the model-checking and the prism tool. The model-checking is very interesting for the evaluation and validation of the probabilistic automaton. They use it to approve the software quality of the three experimental projects. The obtained results are very interesting and of great importance.

Monitoring Social Distancing Using Artificial Intelligence for Fighting COVID-19 Virus Spread

To restrict COVID-19, individuals must remain two meters away from one another in public since public health authorities find this a healthy distance. In this way, the incidence of “social distancing” keeps pace with COVID-19 spread. For this purpose, the proposed solution consists of the development of a tool based on AI technologies which takes as input videos (in real time) from streets and public spaces and gives as output the places where social distancing is not respected. Detected persons who are not respecting social distancing are surrounded with red rectangles and those who respect social distancing with green rectangles. The solution has been tested for the case of videos from the two Holy Mosques in Saudi Arabia: Makkah and Madinah. As a novel contribution compared to existent approaches in the literature, the solution allows the detection of the age, class, and sex of persons not respecting social distancing. Person detection is performed using the Faster RCNN with ResNet-50 as it is the backbone network that is pre-trained with the open source COCO dataset. The obtained results are satisfactory and may be improved by considering more sophisticated cameras, material, and techniques.

redBERT

A natural language processing (NLP) method was used to uncover various issues and sentiments surrounding COVID-19 from social media and get a deeper understanding of fluctuating public opinion in situations of wide-scale panic to guide improved decision making with the help of a sentiment analyser created for the automated extraction of COVID-19-related discussions based on topic modelling. Moreover, the BERT model was used for the sentiment classification of COVID-19 Reddit comments. These findings shed light on the importance of studying trends and using computational techniques to assess the human psyche in times of distress.

Enhancing the Software Clone Detection in BigCloneBench

In the software system, the code snippets that are copied and pasted in the same software or another software result in cloning. The basic cause of cloning is either a programmer‘s constraint or language constraints. An increase in the maintenance cost of software is the major drawback of code clones. So, clone detection techniques are required to remove or refactor the code clone. Recent studies exhibit the abstract syntax tree (AST) captures the structural information of source code appropriately. Many researchers used tree-based convolution for identifying the clone, but this technique has certain drawbacks. Therefore, in this paper, the authors propose an approach that finds the semantic clone through square-based convolution by taking abstract syntax representation of source code. Experimental results show the effectiveness of the approach to the popular BigCloneBench benchmark.

An Open Source Finance System for Stocks Backtesting Trade Strategies

A significant part of pension savings is in the capital market and exposed to market volatility. The COVID-19 pandemic crisis, like the previous crises, damaged the gains achieved in those funds. This paper presents a development of open-source finance system for stocks backtesting trade strategies. The development will be operated by the Python programming language and will implement application user interface. The system will import historical data of stocks from financial web and will produce charts for analysis of the trends in stocks price. Based on technical analysis, it will run trading strategies which will be defined by the user. The system will output the trade orders that should have been executed in retrospect and concluding charts to present the profit and loss that would occur to evaluate the performance of the strategy.

A Hybrid Approach to Identify Code Smell Using Machine Learning Algorithms

Code smell aims to identify bugs that occurred during software development. It is the task of identifying design problems. The significant causes of code smell are complexity in code, violation of programming rules, low modelling, and lack of unit-level testing by the developer. Different open source systems like JEdit, Eclipse, and ArgoUML are evaluated in this work. After collecting the data, the best features are selected using recursive feature elimination (RFE). In this paper, the authors have used different anomaly detection algorithms for efficient recognition of dirty code. The average accuracy value of k-means, GMM, autoencoder, PCA, and Bayesian networks is 98%, 94%, 96%, 89%, and 93%. The k-means clustering algorithm is the most suitable algorithm for code detection. Experimentally, the authors proved that ArgoUML project is having better performance as compared to Eclipse and JEdit projects.