A Machine Learning Approach to Workflow Prioritization

Bollumpally, Niharika R; Evans, Andrew C; Gleave, Scott W; Gromadzki, Alexander R; Learmonth, Gerard P.

doi:10.1109/sieds.2019.8735589

Cited by 1 publication

(11 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also used different names for the algorithms that come under the umbrella of artificial intelligence. Then we identified keywords from the relevant research articles [32], [38], [57], [62], [58], [64], [63], [65]. Finally, we used Boolean OR and AND to create the search strings by combining key terms.…”

Section: ) Search Stringsmentioning

confidence: 99%

“…8) Bollumpally, Evans, Gleave, Gromadzki, and Learmonth (2019) This model presented in [58] employs the XGBoost method, which is a gradient-boosted tree implementation. To reduce over-fitting and maintain high performance, tree boosting groups a large number of slow learners together.…”

Section: ) Gupta and Gupta (2018)mentioning

confidence: 99%

“…All of the surveyed machine learning techniques proposed in [5], [8], [19], [26], [52], [54], [34], [65], and [58] offer scalable solutions, as they can prioritize a large number of software requirements and can be used in a variety of scenarios. Table XXX shows that all techniques can be easily extended and overcome redundancy problems by removing repeating groups of requirements, which leads to efficiency in terms of cost and time.…”

Section: ) Strengths and Weaknesses Of Machine Learning-based Techniq...mentioning

confidence: 99%

“…There is no evidence of efficiency and cost-benefit in [8]. However, the techniques proposed in [19], [52], [54], [65], and [58] are both efficient and cost-effective. Analysts' judgment effort connected with pairwise comparison is kept within reach in these techniques.…”

Section: ) Strengths and Weaknesses Of Machine Learning-based Techniq...mentioning

confidence: 99%

“…For example, fuzzy logic can help in situations where a project has risky and conflicting requirements [44]. Other AI-based techniques that have been applied in prioritizing the requirements include genetic algorithm, simulated annealing, ant colony optimization and machine learning [15], [16], [4], [50], [58], [23], [55], [5].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Systematic Literature Review of AI-Based Software Requirements Prioritization Techniques

Anwar,

Bashir

2023

IEEE Access

View full text Add to dashboard Cite

Software requirements show what the customer desires his software to do. They are the first stepping stone towards a successful software development project. With the increasing complexity of the software due to its size and feature base, it is vital to prioritize the requirements for efficient utilization of development resources. To achieve this, industrial organizations are devising new strategies and improved solutions even with the help of artificial intelligence (AI) tool set. Existing requirements prioritization techniques are human-intensive and suffer from several limitations like overlapping outcomes, scalability problems, time consumption, inaccuracy, and so on. Some of the problems can be solved with the inclusion of artificial intelligence algorithms and strategies. For this several AI-based requirements prioritization techniques have been proposed by applying Genetic Algorithms, Fuzzy Logic, Ant Colony Optimization, and Machine Learning. Literature witnesses some good review studies and surveys on conventional prioritization techniques but there exists none for AI-based techniques that identify not only their strengths but also their weaknesses, advantages of machine learning techniques over other AI-based requirements prioritization techniques, and limitations of applying AI-based techniques in requirements prioritization. This study presents a systematic literature review (SLR) of AI-based requirements prioritization approaches covering 46 papers published from 2000 to 2021. We have given this literature review a new dimension by conducting a parametric analysis of AI-based requirements prioritization techniques and we have identified these parameters after a thorough literature study. Some of the chosen parameters are generic (related to the prioritization process) and some are specific (related to AI techniques). This study has greatly helped us draw a clear line among AI-based techniques to show their domain of application to gain maximum advantage. Our findings will assist researchers, requirement analysts, and other stakeholders in making a wise decision to select the best requirements prioritization technique to gain optimal results.

show abstract