DeepIP

A deep neural network by using a CNN based model and incorporating an attention mechanism for incident prioritization.

Introduction

We propose a deep-learning based approach, called DeepIP (Deep learning based Incident Prioritization), to prioritizing incidents by identifying incidental incidents. DeepTIP.ipynb

Compared Approaches

1. Menzies and Marcus, which first applies the standard text mining method to process textual descriptions in reports, and then uses tf-idf (term frequency-inverse document frequency) to transform the textual description in a report to a vector. Finally, it uses the rule classifier based on entropy and information gain to predict bug severity. Baseline_Rule.ipynb

2. Lamkanfi et al., which applies the standard text mining method to process text descriptions. Then, it counts token frequency and uses the Naive Bayes algorithm to predict bug severity. Baseline_Bayes.ipynb

3. Zhang at al., which applies the standard text mining method and calculates the similarities between a new bug report and historical bug reports using BM25F and LDA (Latent Dirichlet Allocation). According to the similarities, it infers the bug severity based on the top-k nearest neighbors among historical reports (KNN). Baseline_KNN.ipynb

Generality of DeepIP

To investigate the generality of DeepIP, we evaluated the effectiveness of DeepIP on traditional software bug severity prediction using an open-source dataset. Here, we compared DeepIP with the state-of-the-art bug severity prediction approach (Zhang at al.), and used the same Mozilla dataset released in the compared work. Mozilla

Thanks