Comparison Analysis of Facebook's Prophet, Amazon's DeepAR+ and CNN-QR Algorithms for Successful Real-World Sales Forecasting

TL;DR

This paper compares Facebook's Prophet, Amazon's DeepAR+, and CNN-QR algorithms for sales forecasting, analyzing their performance on real-world data with varying sales history lengths to determine their suitability for different sales patterns.

Contribution

It provides a detailed empirical comparison of three popular forecasting algorithms on real sales data, highlighting their strengths depending on data characteristics.

Findings

Prophet performs better with longer sales history and frequent sales.

Amazon's algorithms excel with short history and infrequent sales.

The study offers guidance for selecting forecasting models based on data properties.

Abstract

By successfully solving the problem of forecasting, the processes in the work of various companies are optimized and savings are achieved. In this process, the analysis of time series data is of particular importance. Since the creation of Facebook's Prophet, and Amazon's DeepAR+ and CNN-QR forecasting models, algorithms have attracted a great deal of attention. The paper presents the application and comparison of the above algorithms for sales forecasting in distribution companies. A detailed comparison of the performance of algorithms over real data with different lengths of sales history was made. The results show that Prophet gives better results for items with a longer history and frequent sales, while Amazon's algorithms show superiority for items without a long history and items that are rarely sold.