Context-aware based restaurant recommender system: A prescriptive analytics

Providing recommendations for products or services based on users preferences and current conditions could be more efficient by using a context-based recommender system. It is important and useful to understand the consideration of what should be done by users to visit (prescriptive analytics) based on processed input data optimization. However, discussions and analysis of this system use are still limited. It is noted that prescriptive models can be developed by utilizing or optimizing inputs based on the chosen class rating. In the prediction function, the context-based recommender system can not only be used to predict Good, Neutral, and Bad rating values to produce predictive analytics, but also can be used to optimize input to produce prescriptive analytics. It can be seen that the evaluation of rating predictions using Deep Learning Models showed high accuracy in the performance compared to the Decision Tree and Random Forest. In this model, classification errors were considered the smallest compared to other models. Evaluation of input optimization for prescriptive analytics for class rating predictions showed the highest performance. The research contributes to a better understanding of developing a predictive and prescriptive analytics approach to a context-based recommender system model. © School of Engineering, Taylors University.