Effort estimation is an essential task in a software project as it helps to establish feasible plans for the implementation of a project. It largely influences success or failure of the project. Project planning becomes more efficient with accurate effort estimates, thus providing a number of benefits to the organization. Estimating effort in agile projects has been a challenging task for researchers. Several studies exist in that domain. While some have considered people-related factors, others have catered for project-related factors for estimating effort. Others have adopted machine learning (ML) techniques to produce an accurate estimation. This paper presents a model to estimate and predict effort in a sprint using ML techniques while considering various factors that affect a sprint. The model has been evaluated using various regression algorithms, namely linear regression, K-nearest neighbor, decision tree, polynomial kernel, radius basis function and multi-layer perception (MLP). The model has produced more reliable estimates, with low error values using MLP algorithm.

A predictive model to estimate effort in a sprint using machine learning techniques

Effort Estimation in Agile Software Development: A Exploratory Study of Practitioners' Perspective

• `Agile cost estimation in scrum development. • Cost estimation methods in scrum and limitations. • Suggested a cost estimation model. Agile has been invented to improve and overcome the deficiencies of efficient software development. At present, the agile model is used in software development vastly due to its support to both developers and clients resourcefully. Agile methodology increases the interaction between the developer and client to make the software product defect-free. The agile model is getting to be a well-known life cycle model because of its particular features and most owing is to allow changes at any level of the project from the product owner. However, on other hand, this novel feature is a disadvantage of the agile model due to frequent change requests from the client has increased the cost and time. To overcome cost and time estimation issues different cost estimation techniques are being used in agile development but no one is pertinent for accurate estimation. Therefore, this study has proposed a cost estimation technique. The proposed estimation technique is predictions-based and has different categorizations of projects based on user stories complexities and the developer's expertise. We have applied the suggested technique to ongoing projects to find the results and effectiveness. We have used two projects with different sizes and user stories. Both projects have different modules and developers with different expertise. We have used the proposed estimation technique on projects and done a survey session with the teams. This survey session's main objective is to reveal the statistical findings of the proposed solution. We have designed the 12 hypotheses for statistical analysis. 

https://repositorio.cuc.edu.co/bitstream/11323/9899/1/Prediction%20based%20cost%20estimation%20technique%20in%20agile%20development.pdf

Prediction based cost estimation technique in agile development

This study compares the performance of Pytorch-based Deep Learning, Multiple Perceptron Neural Networks with Multiple Linear Regression in terms of software effort estimations based on function point analysis. This study investigates Adjusted Function Points, Function Point Categories, Industry Sector, and Relative Size. The ISBSG dataset (version 2020/R1) is used as the historical dataset. The effort estimation performance is compared among multiple models by evaluating a prediction level of 0.30 and standardized accuracy. According to the findings, the Multiple Perceptron Neural Network based on Adjusted Function Points combined with Industry Sector predictors yielded 53% and 61% in terms of standardized accuracy and a prediction level of 0.30, respectively. The findings of Pytorch-based Deep Learning are similar to Multiple Perceptron Neural Networks, with even better results than that, with standardized accuracy and a prediction level of 0.30, 72% and 72%, respectively. The results reveal that both the Pytorch-based Deep Learning and Multiple Perceptron model outperformed Multiple Linear Regression and baseline models using the experimental dataset. Furthermore, in the studied dataset, Adjusted Function Points may not contribute to higher accuracy than Function Point Categories.

Comparing Multiple Linear Regression, Deep Learning and Multiple Perceptron for Functional Points Estimation

SEXTAMT: A systematic map to navigate the wide seas of factors affecting expert judgment software estimates

Scrum projects consist of multiple iterations, known as sprints during which user stories are implemented. Estimating user stories of an iteration accurately is important in order to provide clarity and help management to control the project successfully. This leads to the necessity of identifying factors that impact the accuracy of a sprint effort estimation. This paper therefore aims to identify factors and reasons for inaccurate effort estimation of a sprint. For this purpose, a survey is conducted in a scrum environment with professionals with proven agile expertise. The survey is based on 15 completed small-scaled agile projects in a well-known Mauritian company. Results of the survey are summarized in the paper, and recommendations are made.

Melvina Autar Ramessur

Papers

A predictive model to estimate effort in a sprint using machine learning techniques

Factors Affecting Sprint Effort Estimation