Causal aware parameterized quantum stochastic gradient descent for analyzing marketing advertisements and sales forecasting

Nowadays, all marketing companies are focusing on advertisements to maximize their sales and profit. Media mix models are used by advertisers to measure how advertisements would effectively, attract and motivate the consumers to buy their brands. However, due to large number of media mix variables, it is difficult to use these media mix models for producing reliable approximation, extending these approximations to infer causation (not as correlation), and parameterizing these models to handle the complex marketing interactions between consumers. In this work, we develop a novel model integrating causal inference analysis with a parameterized quantum stochastic gradient descent model to assess the relation between the marketing advertisements and sales forecasting. The proposed causal aware PQSGD model is experimented on two different media mix variables datasets. The proposed model outperforms the existing quantum approaches (Quantum linear/lasso regression, Quantum decision tree, Quantum LightGBM, Quantum random forest regressor) by achieving the MSE values of 4.36 for Dataset 1 and 19.24 for Dataset 2.