Sidharth Samal

Dissertation Title: Financial Time-Series Analysis: An Integrated Machine Learning-MCDM Approach. Collaborated with Dr. Rajashree Dash on financial time-series forecasting, integrating MCDM, evolutionary optimization, and ML. Developed a two-stage feature selection framework for financial time-series prediction, reducing error by 40% and data dimensionality by 35%. Created an enhanced neural network weight optimization scheme for financial time-series prediction, reducing error by 30%. Investigated activation functions’ impact, optimizing prediction efficiency for financial trends. Collaborated on dimensionality reduction and optimizing gene feature selection from thousands of gene expressions for cancer detection, improving accuracy. Developed an improved transfer learning framework for cancer detection from brain MRI images, enhancing diagnostic precision.

Assisted in teaching computer science courses, including grading and lab sessions. Mentored undergraduates in research projects, offering technical support and guidance. Collaborated with faculty on research projects, contributing to discussions and experimental design.

Dissertation Title: A TOPSIS Approach of Base Classifier Selection for Crow Search Based Ensemble: In Application to Stock Index Price Movement Prediction. Collaborated with Dr. Rajashree Dash to create an Optimized Weighted MCDM Ensemble Framework, integrating evolutionary optimization. Applied it to predict financial time-series movements, enhancing decision support and introducing a novel weighted voting approach.