Software reusability metrics estimation: Algorithms, models and optimization techniques

Padhy, Neelamadhab; Singh, Rajvinder; Satapathy, Suresh Chandra

doi:10.1016/j.compeleceng.2017.11.022

Cited by 57 publications

(18 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generally, the fast-moving development environment and the tendency for improving software reusability [83,91], means that developers are often looking for easy-to-install frameworks or tools with minimal cognitive overhead, in an affordable and accessible way [29,95]. One way that industry has responded to the interest by non-experts in using ML and AI, is through the provision of cognitive services, algorithmic processes inspired by human cognition, for solving various narrow-AI problems such as sentiment analysis of text, translation of text from one language to another, or visual recognition of objects and concepts depicted in images and video.…”

Section: Auditing Cognitive Servicesmentioning

confidence: 99%

To "See" is to Stereotype

Barlas

Kyriakou

Guest

et al. 2021

Proc. ACM Hum.-Comput. Interact.

View full text Add to dashboard Cite

Machine-learned computer vision algorithms for tagging images are increasingly used by developers and researchers, having become popularized as easy-to-use "cognitive services." Yet these tools struggle with gender recognition, particularly when processing images of women, people of color and non-binary individuals. Socio-technical researchers have cited data bias as a key problem; training datasets often over-represent images of people and contexts that convey social stereotypes. The social psychology literature explains that people learn social stereotypes, in part, by observing others in particular roles and contexts, and can inadvertently learn to associate gender with scenes, occupations and activities. Thus, we study the extent to which image tagging algorithms mimic this phenomenon. We design a controlled experiment, to examine the interdependence between algorithmic recognition of context and the depicted person's gender. In the spirit of auditing to understand machine behaviors, we create a highly controlled dataset of people images, imposed on gender-stereotyped backgrounds. Our methodology is reproducible and our code publicly available. Evaluating five proprietary algorithms, we find that in three, gender inference is hindered when a background is introduced. Of the two that "see" both backgrounds and gender, it is the one whose output is most consistent with human stereotyping processes that is superior in recognizing gender. We discuss the accuracy--fairness trade-off, as well as the importance of auditing black boxes in better understanding this double-edged sword.

show abstract

Section: Auditing Cognitive Servicesmentioning

confidence: 99%

To "See" is to Stereotype

Barlas

Kyriakou

Guest

et al. 2021

Proc. ACM Hum.-Comput. Interact.

View full text Add to dashboard Cite

show abstract

“…Padhy et al [31 ] proposed software metrics algorithms, and their primary constructions for estimating the metrics from the Unified Modeling Language (UML)/class diagrams. They developed the ageing and survivability‐aware (sensitive) reusability optimisation model for object‐oriented software systems.…”

Section: Backgrounds and Related Workmentioning

confidence: 99%

Automatic summarising of user stories in order to be reused in future similar projects

et al. 2020

View full text Add to dashboard Cite

User stories play an important role in agile development systems. In this study, a method of summarising user stories is proposed to reuse them in the future. To enhance the results, quality improvement should be made on user stories. It would help developers build better results, and it may also lead to omitting some essential information. To avoid such issues, user stories are duplicated in two exact similar groups, and quality improvement is made on one set while the other set remains unattained. With the help of a modified bag of words and a verb parser, a collection of keywords and key verbs are extracted for both groups. Afterwards, automatic user stories are made, and then an expert improves them. Next, some experts choose between the results and select the better ones. The result is evaluated by applying different experiments on the framework and prototype implementation on 14 data sets of a user story from industry and a fake data set from Duke University. The result showed 97% of micro F ‐measure and 93% of macro F ‐measure, which are promising. These new user stories can be used as the base user stories in future similar projects.

show abstract

“…Despite the struggle, precise effort estimation is still an open issue. Existing research efforts related to software effort estimation can be broadly divided into two groups: (i) non-algorithmic, and (ii) algorithmic [7]. Techniques based on non-algorithmic models [8,9] include analogy-based estimation, expert judgment, Fuzzy Logic (FL) [14], learning-based techniques (Artificial Neural Network (ANN)) [9,40], Machine Learning (ML) [11,12], Case-Based Reasoning (CBR) [13].…”

Section: Introductionmentioning

confidence: 99%

Metaheuristic Algorithms in Optimizing Deep Neural Network Model for Software Effort Estimation

et al. 2021

View full text Add to dashboard Cite

Effort estimation is the most critical activity for the success of overall solution delivery in software engineering projects. In this context, the paper's main contributions to the literature on software effort estimation are twofold. First, this paper examines the application of meta-heuristic algorithms to have a logical and acceptable parametric model for software effort estimation. Secondly, to unravel the benefits of nature-inspired meta-heuristic algorithms usage in optimizing Deep Learning (DL) architectures for software effort estimation, this paper presents a Deep Neural Network (DNN) model for software effort estimation based on meta-heuristic algorithms. In this paper, Grey Wolf Optimizer (GWO) and StrawBerry (SB) meta-heuristic algorithms are applied for having a logical and acceptable parametric model for software effort estimation. To validate the performances of these two algorithms, a set of nine benchmark functions having wide dimensions is applied. Results from GWO and SB algorithms are compared with five other meta-heuristic algorithms used in literature for software effort estimation. Experimental results showed that the GWO has comprehensive superiority in terms of accuracy in estimation. The proposed DNN model (GWDNNSB) using meta-heuristic algorithms for initial weights and learning rate selection, produced better results compared to existing work on using DNN for software effort estimation.

show abstract

Software reusability metrics estimation: Algorithms, models and optimization techniques

Cited by 57 publications

References 4 publications

To "See" is to Stereotype

To "See" is to Stereotype

Automatic summarising of user stories in order to be reused in future similar projects

Metaheuristic Algorithms in Optimizing Deep Neural Network Model for Software Effort Estimation

Contact Info

Product

Resources

About