Demand Forecasting in Artificial Intelligence
Demand Forecasting model was demonstrated by CellStrat AI Lab members Shreyas S.K., Natarajan Lalgudi, Anupam Ranjan & Dr. Purnendu Sekhar Das in CellStrat AI Conclave on 8th. February’20 in Bangalore.
Demand Forecasting is the activity of estimating the quantity of a product or service the consumer may purchase. The retail industry has been facing challenges of predicting the demand accurately which results in enormous loss in revenue. As per statistical data, retailers loose more than 50% customers due to stock out items.
As per Gartner Group and AMR Research, accurate demand forecasting results in massive amount of savings to the organizations and results in higher customer satisfaction.
Demand forecasting traditionally been performed through statistical techniques like ARIMA, Auto Regression and with recent advancements in machine/deep learning, more sophisticated and accurate prediction has been possible.’
Statistical techniques like Auto Regression, ARIMA has been traditionally used. Facebook released Prophet, their time series library in late 2018.
XG Boost and Light GBM are popular algorithms for time series prediction
LSTM gives excellent result for time series analysis.
They presented two solutions to demand forecasting
- Facebook Prophet based time series forecasting
- Machine Learning based forecasting using XG Boost
View complete Poster here.