Linear Support Vector Regression for Next-Day Index Trades

Summary

This example applies a linear support vector regression model to daily closes of the CSI 300 index. It constructs features from the current close and the preceding twenty daily closes, then fits the model to predict the following day’s close. The historical observations are divided into a training segment and a later test segment, although the displayed trading and cumulative-return calculations are shown for the training predictions. A long position is taken when the predicted next close exceeds the next day’s open by a configured trigger; the position is evaluated as buying at that open and selling at that day’s close.

The code compares compounded strategy returns with a close-to-close index benchmark, but the document provides no plotted values or performance statistics, so it does not demonstrate predictive or trading efficacy. Its illustrative calculations do not describe transaction costs, slippage, or a full out-of-sample evaluation. The timing and availability of the next open should also be handled explicitly when turning the prediction rule into executable orders.

Key ideas

• The model uses a linear support vector regressor to estimate the next daily close from recent closes.
• The example creates lagged close features spanning twenty prior sessions.
• It enters a long intraday trade when the predicted close clears the next open by a trigger threshold.
• The code defines a compounded return series and a close-to-close benchmark but reports no results.
• Trading costs and precise signal execution timing are not addressed.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.