Sumathi Chakravarthy, Badri Narayanan G.
To our knowledge, recommendation systems in the current literature do not account for preferences in terms of responsiveness to prices and quality. The major goal of this project is to fill this gap in the literature by employing the economic concept of elasticity with respect to prices and quality. Elasticity of demand with respect to any variable (price, income, quality, etc.) is the ratio of relative changes in quantity to those in the variable in question. Collaborative and content-based filtering methods, as well as their hybrid systems, take into account various characteristics of past purchases, but we will attempt to venture a step ahead: incorporate the behavior of the customer ‘at margin’, i.e., compute the degree of responsiveness to changes in prices and ratings. The algorithm involves two stages: some recommendations are short-listed using existing methods in the first stage and they are further fine-tuned using elasticity in the second stage, into “buy” and “not buy” categories. We also predict the percentage of changes in purchases of items, in response to changes in prices and quality. We illustrate the usefulness of this methodology using the movie rating dataset.