Predicting Software Defects Based on Cognitive Error Theories

Abstract

As the primary cause of software defects, human error is the key to understanding and perhaps to predicting and preventing software defects. However, little research has been done to forecast software defects based on defects' cognitive nature. This paper proposes an idea for predicting software defects by applying the current scientific understanding of human error mechanisms. This new prediction method is based on the main causal mechanism underlying software defects: an error-prone scenario triggers a sequence of human error modes. Preliminary evidence for supporting this idea is presented.

Requirements‐related fault prevention during the transformation from formal specifications to programs

“…Therefore, the root causes are divided into two categories: human and specification. Human errors are the identified primary causes for these faults, which are consistent with the existing investigation [10,12,39,40]. The programmers may not understand the specification comprehensively and do not capture the implicit dependence relationships among the modules, the type and variable declarations, and process specifications in the SOFL formal specification and they make the undefined feature fault reported by IDE in Java programming.…”

“…The identified root causes are generally classified into four categories: human, requirements, environment, and tool [10,60]. Human-related errors are identified as the primary root causes [4,10,39,40]. In particular, we find that the programmers often have difficulty understanding formal specifications consisting of complex predicate logic formulas and knowing how to transform such specifications to programs.…”

Software Defect Prediction Based on Elman Neural Network and Cuckoo Search Algorithm

“…Felix and Lee [13] presented the application of an integrated machine-learning approach based on regression models constructed from these predictor variables, which enhanced the effectiveness of software development activities. In 2018, Huang and Strigini [14] applied the scientific understanding of human error mechanisms to predict software defects.…”

“…where x 1 and x 2 are scores of approaches 1 and 2, respectively. In (14), n 1 and n 2 mean the sizes of the sample groups of the above approaches. e cardinality symbol # denotes counting.…”

Targeted Code Inspection based on Human Errors

“…Huang and Liu [1] first proposed the area "Software Fault Defense based on Human Errors" to systematically reduce software defects based on a deep understanding of the human errors of software practitioners. Huang et al has carried out in-depth studies on a variety of topics, such as defect prevention based on human errors [2,3], software diversity design based on human errors [4], and defect prediction based on human errors [5,6]. Anu and Walia et al have been focusing on using human error taxonomy to improve requirement quality [7].…”