|
Title:
|
SOFTWARE EFFORT ESTIMATION
WITH METAHEURISTIC OPTIMIZED ENSEMBLE
FOR SOFTWARE CROWDSOURCING |
|
Author(s):
|
Anum Yasmin and Wasi Haider Butt |
|
ISBN:
|
978-989-8704-48-1 |
|
Editors:
|
Miguel Baptista Nunes, Pedro Isaías and Philip Powell |
|
Year:
|
2023 |
|
Edition:
|
Single |
|
Keywords:
|
Software Effort Estimation, Crowdsourcing, Machine Learning, Ensemble Effort Estimation, Metaheuristic Optimization |
|
Type:
|
Full Paper |
|
First Page:
|
161 |
|
Last Page:
|
171 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Software crowdsourcing (SWCS) is rapidly growing from past decade due to its flexible work environment. Software
effort estimation (SEE) is already renowned field in traditional software engineering, utilized in preliminary resource
planning, budget, and time. SWCS platform can suffer from schedule, cost, and human resource uncertainty, which can
be attained by estimating effort consumed on crowdsourced tasks. With more advent in SEE, ensemble effort estimation
(EEE) is emerged providing unbiased results across different datasets. Recently, intelligent methods such as metaheuristic
algorithms are used for ensemble by assigning optimal weights. This study aims to utilize prediction accuracy of EEE and
metaheuristics on SWCS tasks to established accurate effort estimation model. In this work, ensembles are created using
high predictive solo machine learning algorithms (RF, SVM, NeuralNet), whose weights are optimized with MO.
TopCoder is selected as target SWCS platform, and this is first work to utilize TopCoder Design category tasks and
contributes in dataset formation with relevant crowdsourced designing features. Results of proposed scheme clearly show
that Metaheuristic-weight learning is giving more accurate ensembles with approximately 60% performance
improvement compared to solo ML and other EEE techniques, proving it more suitable SEE technique for SWCS. |
|
|
|
|
|
|
