Search

Up to now, several metrics have been determined and arisen in order to measure and evaluate software developers' performance: Productivity, Engagement, Attention to Quality, Code Based Knowledge, Skills and Management, Adherence to Coding General Structure and Techniques, Learning and Skills, Personal Responsibility, etc. However, there is not any universally accepted a set of metrics or a methodology to measure and evaluate software developers' performance. Three main reasons of this situation which may be thought; Firstly, each part of software creation is unique. There is no compelling reason to assemble two times the same parts of software as it might be duplicated by copying it. This makes it truly difficult to make a formal and thorough correlation between two parts of software based on both qualitative and quantitative approach. Secondly, the current technology is something that changes at a truly fast pace (phase). So, as a natural result of this situation, each time a methodology in respect to a certain wave of technology or technology-based metrics is dependable enough, it is for the most part as of recently old and obsolescence in time. Thirdly, there is a gigantic zone for innovativeness in discovering the diverse answers for a unique issue. For instance, calculating coding lines that software developers write for the same problem means size of the solution, not the problem. This case shows us difficulty of turning or converting tangible data to the metric in software developers' performance. In this study, the researches, which have been done before about performance criteria, and the case studies, which have been published before about performance criteria, have been observed, analyzed and evaluated (in total 18 separate research and study). In the light of these consequences, a measurement and evaluation criteria set about the software developers' performance have been created and centered into "Software Engineering". [ABSTRACT FROM AUTHOR]

Multi-instance classification is a supervised learning technique which aims to predict the labels of the bags that consists of unlabeled instances with multi-feature vectors. In the last decade, ensemble learning has become one of the most popular machine learning methods since it generally provides more accurate classification results than the individual classifiers. Considering this motivation, the study presented in this paper combines these two techniques to benefit from their capabilities together. This paper proposes an ensemble-based multi-instance classification (EMIC) approach which combines a set of individual multiinstance classifiers to obtain final class label using a voting mechanism. This study also compares the suggested EMIC approach with the traditional multi-instance classification methods in terms of classification accuracy. In the experimental studies, the EMIC approach was tested on various real-world and well-known multi-instance datasets. The results indicate that the proposed ensemble approach generally achieves higher classification performance than the individual multi-instance classification algorithms. [ABSTRACT FROM AUTHOR]