Authenticating Users In Real World Applications using Multi Modal Biometric System For Smartphone’s Based Hidden Markov Model Compared With K Nearest Neighbor Algorithm

The overarching goal of this research is to improve Smartphone user identification in practical settings by using multi-model biometric systems. A dataset was retrieved from the Kaggle library for this specific research topic. Hidden Markov model with K-nearest-neighbor in practical settings, themethod is used to verify users’ identities using Smartphone multimodal biometric systems, with distinct training and testingphases. Around 85% of the time, a G-power test will provide theexpected result with the default settings of =0.05 and power =0.85. The significance level was determined to be 0.014 with a p-value lower than 0.05. The Hidden Markov model outperforms the K closest neighbor algorithm (96.4770%), which was previously ranked at 93.8980%. Finally, when comparing accuracy between the Hidden Markov model and theK-Nearest Neighbor Algorithm, the former comes out on top statistically.