Modelo de propensión a compra de clientes = Customer propensity to buy model

Resumen

Customer purchase forecasting aims to predict future customer purchases. The resulting data is of great importance to perform future business activities by designing strategies that match consumer preferences. To obtain accurate predictions of customer purchases, this paper develops a machine learning framework based on historical data. Machine learning techniques are applied to predict which customers have a higher propensity to purchase and ranking techniques are applied to prioritize which customers to act on first based on those that provide the highest value, based on RFM (Recency, Frequency, Monetary) analysis. Due to the high competition and the numerous loyalty plans that exist in the market, there is a need to determine which are the most valuable consumers for the online store and to know how likely they are to make a repeat purchase. To do this, consumers registered in the loyalty program of an online store are analyzed. The loyalty plan will measure purchase frequency, spending, how current the customer is, as well as different sociodemographic measures. First, the data is explored to get an idea of what kind of information is available and how it can be used to predict purchases. Then different machine learning algorithms are tested on the data and their performance is compared. After this, the parameters of the chosen algorithm are adjusted to further improve its performance. Finally, a dataset is adopted to test the feasibility of the proposed prediction framework. Experimental results and comparative analysis verify the validity of the proposed model.