isiforecast

Isiforecast: Promoting innovative and smart economic transformation

Funding body: European Union 60% + Normandy Region 40%
Total funding: €81,000
Project period: 1 November 2023 to 30 April 2026 (30 months)
Principal Investigators: Alexandre Berred and Cyrille Bertelle

Partners

• LITIS – Laboratory of Computer Science, Information Processing and Systems (ULHN);
• CIRCOE, Logistics Consultancy and Innovation, Normandy, Le Havre;
• Artyfêtes, Designer of collections of party supplies for professionals.

Project description

Effective stock management and sales forecasting are major challenges for businesses, regardless of their sector. Failure to anticipate demand accurately can lead to costly overstocking, stock-outs or a loss of business opportunities. Against this backdrop, the ISIFORECAST project aims to provide an intelligent and flexible platform that improves the accuracy of sales forecasts, whilst remaining accessible to users with a variety of profiles. ISIFORECAST combines traditional forecasting methods (such as SARIMAX and Prophet) with advanced deep learning models (N-BEATS, TCN, LSTM and hybrid approaches) to capture the complex temporal dynamics of sales time series. The project also incorporates additional features such as product clustering, product life-cycle analysis and the management of the effects of stock-outs, thereby providing a comprehensive and operational view of forecasts.

Isiforecast is now being rolled out by Circoe: Isiforecast website