Prioritizing software maintenance plan by analyzing user feedback

In order to adapt with evolving requirements and perform continuous software maintenance, it is essential for the software developers to understand the content of user feedback. User feedback such as bug report could provide so much information regarding the product from user’s point of view, especially parts that need improvements. However, it is often difficult to read all the feedback for products with enormous number of users as manually reading and analyzing each feedback could take too much time and effort. This research aims to develop a model for automatic feedback classification by implementing Support Vector Machine for the classifier’s algorithm and Chi-square method for feature selection. The model is developed using Python programming language and is then evaluated under different scenarios in order to measure its performance. Using a ratio of training and testing set of 80:20, our model achieved 77% accuracy, 50% precision, 55% recall, and 73% F1-score with 6.63 critical value and C=100 and gamma 0.001 as the SVM hyperparameters.

Download Full-text

Optimization of Condition-Based Maintenance for Industrial Gas Turbines: Requirements and Results

Volume 5: Microturbines and Small Turbomachinery; Oil and Gas Applications ◽

10.1115/gt2009-59935 ◽

2009 ◽

Cited By ~ 1

Author(s):

Markus Bohlin ◽

Mathias Wa¨rja ◽

Anders Holst ◽

Pontus Slottner ◽

Kivanc Doganay

Keyword(s):

Gas Turbine ◽

Preventive Maintenance ◽

Gas Turbines ◽

Software Maintenance ◽

Planning Process ◽

Economic Effects ◽

Time Interval ◽

Maintenance Planning ◽

Custom Made ◽

Maintenance Plan

In oil and gas applications, the careful planning and execution of preventive maintenance is important due to the high costs associated with shutdown of critical equipment. Optimization and lifetime management for equipment such as gas turbines is therefore crucial in order to achieve high availability and reliability. In this paper, a novel condition-based gas turbine maintenance strategy is described and evaluated. Using custom-made gas turbine maintenance planning software, maintenance is repeatedly reoptimized to fit into the time intervals where production losses are least costly and result in the lowest possible impact. The strategy focuses on accurate online lifetime estimates for gas turbine components, where algorithms predicting future maintenance requirements are used to produce maintenance deadlines. This ensures that the gas turbines are maintained in accordance with the conditions on site. To show the feasibility and economic effects of a customer-adapted maintenance planning process, the maintenance plan for a gas turbine used in a real-world scenario is optimized using a combinatorial optimization algorithm and input from gas turbine operation data, maintenance schedules and operator requirements. The approach was validated through the inspection of a reference gas turbine after a predetermined time interval. It is shown that savings may be substantial compared to a traditional preventive maintenance plan. In the evaluation, typical cost reductions range from 25 to 65%. The calculated availability increase in practice is estimated to range from 0.5 to 1%. In addition, downtime reductions of approximately 12% are expected, due solely to improved planning. This indicates significant improvements.

Download Full-text

Feedback2Code: A Deep Learning Approach to Identifying User-Feedback-Related Source Code Files

International Journal of Software Engineering and Knowledge Engineering ◽

10.1142/s0218194020500011 ◽

2020 ◽

Vol 30 (01) ◽

pp. 1-22

Author(s):

Shuhan Yan ◽

Tianjiao Du ◽

Beijun Shen ◽

Yuting Chen ◽

Zhilei Ren

Keyword(s):

Deep Learning ◽

Open Source ◽

Software Maintenance ◽

Source Code ◽

User Feedback ◽

Software Systems ◽

Learning Approach ◽

Software Products ◽

Process Feedback ◽

Core Idea

Users frequently raise feedback when using software products. Feedback from users regarding their experiences and expectations and software defects they found adds values to software maintenance and evolution — software managers collect user feedback and then dispatch feedback issues that developers (and/or maintainers) need to track and process. Feedback tracking is often supported by open source platforms and collaborative software systems. Meanwhile, there still exists a gap between feedback issues and source code: since user feedback is usually informal and arbitrary, engineers have to spend much effort on comprehending issues and identifying which source code files need to be improved or fixed. This paper introduces a deep learning approach, Feedback2Code , which facilitates identification of user-feedback-related source code files. The core idea is to (1) explore latent semantics of user feedback and source code using several deep learning techniques such as Multi-Layer Perceptron (MLP), Convolutional Neutral Network (CNN) and skip-gram and (2) establish a multi-correlation model to explore linkages between feedback issues and source code files. Given a feedback issue, the linkages then allow engineers to identify source code files that are highly relevant to the issue. We have implemented Feedback2Code and evaluated it against ChangeAdvisor (a state-of-the-art approach) on 24 open source projects. The evaluation results clearly show the strength of Feedback2Code : for 103793 feedback issues, Feedback2Code successfully established 101190 feedback-code linkages and achieved a precision that is [Formula: see text] higher than that of ChangeAdvisor . Feedback2Code also achieved an MRR and an MAP that are [Formula: see text] and [Formula: see text] higher than those of ChangeAdvisor , respectively. Furthermore, we also found that a Feedback2Code -trained model can be easily transferred, allowing feedback-code linkages to be established in new projects with a little history data.

Download Full-text