Design of a Machine Learning based Data Analytics Engine to Improve Decision Making Processes

Analytical Hierarchy Processing (AHP) assists in estimation of input-specific weights for improving system-level decision making operations. The process initially decides a goal, and then uses different criterion & alternatives to achieve it. But original AHP model is not efficient for large-scale systems, which limits its capability under real-time use cases. To overcome this issue, various Machine Learning based techniques are proposed, and each of them has their own set of limitations. Techniques with higher efficiency also showcase high complexity, while simplistic optimization techniques have limited performance levels. To overcome these limitations, this text proposes design of an efficient AHP model with low complexity via integration of (GWO), and iterative learning (IL) operations. The proposed model initially performs context-based correlation between input parameter sets. These correlation values are used to modify the AHP decisions, which are tuned by GWO process. Results of Grey Wolf optimization this GWO process are optimized via IL based tuning, which incrementally modifies the correlation values for optimum efficiency levels. Due to which the proposed model showcases low complexity, while maintaining higher efficiency when compared with the original AHP process. The proposed model was evaluated on multiple datasets including housing, image processing, and employment scenarios. Based on this evaluation it was observed that the proposed model showcased 8.5% higher accuracy, 6.4% higher precision, with 4.5% 1ower delay when compared with original AHP and standard AHP-optimization techniques. Due to this improvement, the proposed model is capable of deployment for a wide variety of real-time scenarios.