Finally, we combine these two values to calculate the technical debt principal. Then, for each instance, we statically analyse the source code to calculate the exact number of lines of code creating the smell. To do so, we adopt novel techniques from Information Retrieval to train a learning-to-rank machine learning model that estimates the severity of an architectural smell and ensure the transparency of the predictions. Our approach can estimate the amount of technical debt principal generated by a single architectural smell instance.
In this paper, we propose a novel approach to estimate architectural technical debt principal based on machine learning and architectural smells to address such shortcomings. Moreover, a recent study has shown that many of the current approaches suffer from certain shortcomings, such as relying on hand-picked thresholds. The current literature contains an array of approaches to estimate TD principal, however, only a few of them focus specifically on architectural TD, and none of these are fully automated, freely available, and thoroughly validated. Bug classification techniques have already been applied in the bug severity prediction and bug triaging areas but not in the CQC applications.Ī key aspect of technical debt (TD) management is the ability to measure the amount of principal accumulated in a system. Code quality control (CQC) is a vital task over bug-fixing process to delay the code deterioration by refactoring activities.
In contrast, the simpler one-phased method performed comparable to the two-phased method in low-predictable products. The two-phased method outperformed the one-phased method in products involving highly predictable classes’ time series. The two-phased method was able to identify unpredictable bug reports by analyzing the time series of its target classes. Using four large-size open-source Apache products, it was observed that the two-phased approach could reach 78% prediction accuracy.
The former makes the predictions using a convolutional neural network and the latter is based on the bug localization algorithms (as the first step) and time series prediction techniques (as the second step). One-phased and two-phased bug classification models are proposed in this paper. The purpose is to alert the project manager to the presumptive quality critical bugs (QCB) as soon as the bug reports are recorded in the issue tracking system (ITS) and help treat them more carefully by assigning those to the more experienced developers to be fixed or by prioritizing those QCBs in the quality control list. Gradual code deterioration is the result of such practice and brings about hard to maintain code by affecting the code quality adversely. Software maintenance phase involves successive code changes due to the reported bugs causing the emergence of bad smells in the code.
0 kommentar(er)
