• S Durga Devi
• R Rama Krishna Reddy
• S. Revanth Reddy
• K. Spandana

: Machine learning models are constructed using training data, which is biased and derived from human experience. Humans exhibit cognitive bias in their thoughts and conduct, which is reflected in the data acquired. The best machine learning models are said to mimic humans’ cognitive ability, and thus such models are inclined towards bias. However, detecting and evaluating bias is important for better explainable models. In this paper, we aim to identify bias in learning models in relation to humans’ cognitive bias and propose a novel technique for detection and evaluation of machine learning bias. In this paper, we try to detect the bias in the dataset. In the deployed dataset, the potentially biased attributes are observed. To detect bias, first we select some common biased attributes and then we use the concept of alternation function to swap the values of PBA’s and evaluate the impact on prediction using KL Divergence. In this paper, we compare the KL divergence value obtained from different models used for training the dataset and predicting the output (average income).

Bias in ML, cognitive Bias, bias detection and evaluation

"DETECTION AND EVALUATION OF MACHINE LEARNING BIAS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.9, Issue 6, page no.j668-j683, June-2022, Available :http://www.jetir.org/papers/JETIR2206985.pdf

"DETECTION AND EVALUATION OF MACHINE LEARNING BIAS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.9, Issue 6, page no. ppj668-j683, June-2022, Available at : http://www.jetir.org/papers/JETIR2206985.pdf