Integrating Multi-Criteria Decision Making and Random Forest Framework for Product Advisory System

Complexity in decision making of product evaluation, healthcare, finance, and policy areas often refers to trade-offs between a variety of contradictory criteria. The training approaches that have historically always been used in one form or another can rarely capture such complexity and so outcomes that are not accurate representations of what should be given importance in the real world can occur. This study will be focused on creating a general-purpose decision support system that is based on an extensive range of multi-criteria solutions that can be used to make decisions and machine learning to become more reliable and flexible. The model is a hybrid of the Weighted Sum Model, Technique of Order Preference by similarity to ideal solution, Preference Ranking Organization Method of enrichment Evaluation, Analytic hierarchy Process, elimination and choice Expressing Reality, Multi objective optimization by ratio Analysis, complex proportional Assessment, Analytic Network Process and the grey Relational Analysis. Meanwhile, a Random Forest learning model is introduced to merge latent connections with the data and associate the rankings come by the decision-making approaches. Practical testing of the architecture based on product sales records proves that the proposed system can combine user defined preferences with predictive ratings to generate stable and understandable recommendations. The findings suggest that this hybrid model is not restrictive to product choice, but it could be extended in terms of application in some other aspects like service evaluation, patient treatment planning, investment ranking, and policy evaluation. The study introduces a generalizable and customizable approach that takes the decision support beyond the classical models and analytics based on data.