Research Papers - Dept of Information Technology

Permanent URI for this collectionhttps://rda.sliit.lk/handle/123456789/593

Browse

Filters

2023 - 20262

Advanced Search

Filter by

Settings

Author: search.filters.author.Haddela, P.SStart date: 2023

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    PublicationOpen Access
    EuqAud: Detecting Gender Bias in Audio Datasets Using Polynomial Regression-Based Metric
    (Institute of Electrical and Electronics Engineers Inc., 2026) Jayawardena, S; Haddela, P.S; Shyamalee, T; Ekanayake, A; Mudalige, T; Dhanawardhana, I
    With the growing adoption of audio based AI systems in high-stakes domains such as healthcare, law enforcement, and social media, ensuring fairness particularly regarding gender bias has become critically important. While prior work on fairness has predominantly focused on disparities in model performance, bias inherent in training datasets remains underexplored. To address this gap, we propose EuqAud, a novel, pre-trained and traceable fairness metric that quantifies gender bias in audio datasets using raw acoustic features such as pitch, energy, amplitude, and voice activity. Unlike methods dependent on demographic labels such as race, age or language, EuqAud is designed to be demographic and language agnostic, enhancing its applicability across diverse contexts. The score is computed using an equation derived from polynomial regression with L2 regularization (Ridge regression), yielding robust and generalizable outputs. It spans a range from −10 to 10, where 0 denotes neutral, positive scores indicate male dominant bias, and negative scores reflect female dominant bias. For clarity, bias severity is categorized into three tiers: Neutral (EuqAud < 2), Moderate Bias (2 ≤ EuqAud ≤ 6), and Strong Bias (EuqAud > 6). Evaluation across multiple datasets demonstrates high predictive performance, with R2 values between 0.95 and 0.99. By focusing on dataset level bias rather than model outcomes, EuqAud offers a scalable and rigorous solution for advancing fairness in audio-based AI systems.
  • Thumbnail Image
    PublicationEmbargo
    CricSquad: A System to Recommend Ideal Players to a Particular Match and Predict the Outcome of the Match
    (IEEE, 2023-06-12) Lekamge, E. L.; Wickramasinghe, K. R.; Gamage, S. E.; Thennakoon, T. M. K. L.; Haddela, P.S; Senaratne, S
    Selection of the cricket squad plays a very important role in the outcome of the match. This work is about selecting ideal players for a cricket match and predicting the outcome of the match according to the selected cricket team. A cricket squad consist of around 15 to 16 players, with different expertise in batting, bowling, fielding. To select players for the squad, points were calculated using a statistical approach considering player’s overall career data. And then for the further use of selecting players for the squad next match performance of each and every player were predicted using Machine Learning techniques. Association rule mining was used to find frequent winning player combinations with day/night, home/away, batting first/second, against different opponent combinations. Finally calculate points for each player in both teams, then predict the outcome of the match with classification algorithms by considering the calculated total points of each team and other factors such as toss outcome, batting inning, day night conditions and venue. As for the results, XG boost regressor has produced the highest R2 score of 0.92 for batsman runs prediction model while random forest regressor has produced the highest R2 score of 0.66 for bowler wickets prediction model. The Gradient Boost Classifier predicted the Outcome of a match with the highest accuracy of 0.92 while the K Nearest Neighbor achieved the lowest accuracy of 0.82 score.