Daily Volume Forecasting using High Frequency Predictors

L.G.M. Alvim, C.N. dos Santos, and R.L. Milidiú (Brazil)


Finance, Volume Forecasting, Machine Learning, PLS, SVR.


Daily volume is an important feature when it comes to financial market structure. Effective daily volume forecasting can help areas such as portfolio management and algorithm trading. Intraday updates of daily volume forecasts can explore high frequency data to provide more accurate forecasts. Previous work on daily volume forecasting usually use Bayesian methods. We approach the problem of daily volume forecasting using intraday information. Here, forecasting is accomplished by two machine learning based predictors: Support Vector Regression (SVR) and Partial Least Squares (PLS). We empirically test our method us ing the top nine high liquidity Bovespa traded stocks. The results indicate that SVR and PLS predictors provide accurate forecasts. Moreover, the forecasting accuracy improves throughout the day as more intraday information is available.

Important Links:

Go Back