IHS Markit deploys machine learning to expand dividend forecasting tool

IHS Markit’s dividend forecasting service has been expanded with the help of advanced analytics and machine learning models.

IHS Markit has implemented advanced analytics and machine learning technology to double the number of companies covered in its dividend forecasting tool.

The expansion means that 28,000 stocks are now covered in more than 90 markets globally, providing up to five years of dividend forecast data in each stock. The tool processes a range of inputs and proxies such as company guidance, historical patterns, industry and peer group trends, and consensus forecasts.

“Using machine learning, advanced statistical modelling and time series analysis techniques to uncover patterns in multiple datasets helps us dramatically expand coverage and also retain the analytical rigor our dividend forecasting team has developed over many years,” said Yaacov Mutnikas, chief data scientist and chief technology officer at IHS Markit.

The new model to expand coverage of the tool was developed with IHS Markit’s data science team, using machine learning technology to apply advanced methods of forecasting at increased scale and scope to process datasets.

Alessandro Ferretti, head of dividend products at IHS Markit, also commented that the move will see IHS Markit’s 40 expert forecasters’ fundamental analysis boosted with augmented intelligence and quantitative insights produced by the new machine learning-based model.

“Combining cutting edge analytics with our more than 15 years of Dividend Forecasting expertise offers index providers, derivatives desks, asset managers and hedge funds access to the broadest set of dividend forecasts available for large-, mid- and small-cap stocks globally, including frontier markets,” Ferretti said.