Time series forecasting in Odoo

Details

Supervisors

Legay, Axel

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en informatique, à finalité spécialisée

Abstract

Time series forecasting is a popular branch in data analysis that allows to detect possible trends or seasonal patterns in an ordered list of values and enables to predict future values. This thesis presents the implementation of an automatic time series forecasting algorithm in the Odoo software and, more specifically, an automatic exponential smoothing forecasting algorithm. Automatic means that the model is trained automatically and the parameters of the model are automatically computed. The main goal of this algorithm is to predict future demand based on recorded sales, and with this to extend a feature in Odoo that schedules the future production or replenishment of a company’s products. The thesis contains a theoretical review of time series and exponential smoothing methods, a description of the algorithm, and validation tests to show the accuracy of the predictions.