Principal Component Analysis of Lack of Cohesion in Methods (LCOM) metrics

In this report, we study the Lack of Cohesion in Methods (LCOM) metric for an object-oriented system and examine the suitability of eight variations of this metric through a principal component analysis. 1. Introduction One concern in software engineering is how high-quality software can be produced with predictable costs and time. Software metrics provide a quantitative means to predict the software development process and evaluate the quality of the software products. Several software metrics have been proposed to measure the complexity in the procedural paradigm. Some of the metrics which are frequently used in the procedural paradigm are McCabe's cyclomatic complexity metric [1] and Halstead's software science metric [2]. The object-oriented programming paradigm is often claimed to allow a faster development pace and higher quality of software. However, software metrics are less well studied in the object-oriented paradigm. A small number of metrics have been proposed to measur..