Mitsui Chemicals teams up with NEC and dotData to trial AI-based price change forecasting for market-sensitive products

2 Mins read
price change forecasting

Mitsui Chemicals, Inc. has partnered with NEC Corporation to test an AI-based system to predict changes in the prices of specific Mitsui Chemicals products that are sensitive to market conditions. Based on the trial results, Mitsui Chemicals aims to work towards the complete introduction of an AI-powered demand forecasting system to reduce losses from price fluctuations and increasing profits by appropriate procurement, production, & sales for their business.

The trial was executed through the dotData AI software platform, developed by dotData, Inc., enabling automation of all data science processes.

“Amid increasingly intense global competition, Mitsui Chemicals intends to become more competitive via the digital transformation of demand forecasting, tapping into advanced automated machine learning technology. This move will speed up our efforts to achieve corporate transformation focused on a customer-driven business model and solving social issues by reducing procurement costs, cutting lead times and reducing CO2 emissions through the optimization of distribution,” said Masao Sambe, executive officer, Digital Transformation Division, Mitsui Chemicals.

Mitsui Chemicals has traditionally estimated trends in product demand based on past price and profit trends, and from exchange rates and other data collected weekly, mainly based on the experience and knowledge of the company’s business staff. However, with the progress in globalization in recent years, it has become difficult to predict sudden changes in market needs, which has made it necessary to control not only output, but also procurement costs of raw materials and quantities.

“Through our partnership with dotData, NEC has supported numerous clients to date in achieving digital transformations. The combination of dotData’s market-leading technology and NEC’s consulting capabilities will not only support Mitsui Chemicals’ efforts to achieve innovative change but also contribute to resolving social issues through this initiative,” said MORI Hideto, Executive Director, NEC.

During the trial, Mitsui Chemicals chose various company products to test. These products were those that were most likely to be affected by difficult-to-forecast market conditions.

The NEC-supplied dotData platform was used over the last few years to analyze a variety of data, including daily and weekly inventory data for products, as well as plant operation rates and sales figures. dotData automatically extracted the most valid of a myriad of possible features to create a high-precision price forecasting model.

While some areas still require ongoing improvement, testing has shown how dotData can be used to make highly accurate forecasting of the next month’s price for a given product, allowing Mitsui Chemicals to more reasonably price products based on trends in the market, formulate better sales plans, and execute procurement and production according to those plans. Mitsui Chemicals estimates that the adoption of the platform will reduce inventory worth several hundred million yen.

“We at dotData are excited to provide our high-precision AI-based forecasting and business insights to assist Mitsui Chemicals in its digital transformation and its efforts to solve the challenges faced by society. We look forward to achieving further progress with Mitsui Chemicals and NEC as partners in this collaboration,” said FUJIMAKI Ryohei, Ph.D., CEO, dotData.

This trial has enabled the company to get new business insights that were not readily apparent to human analysts, such as the identification of a correlation between any change in X and the price of Y.

Mitsui Chemicals established the Digital Transformation Division in April 2021. Focused particularly on business operations and supply chains, the Digital Transformation Division is working to speed up digital transformation efforts company-wide.

Read Next: NEC iris recognition technology ranks first in NIST accuracy testing

1 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

90 ÷ = 10