Intraday Trading Volume and Non-Negative Matrix Factorization.

The intraday trading volume of a security is the total amount of traded contracts distributed over the day. Consequently, the intraday trading volume captures part of the intraday trading activity and represents a proxy for the intraday liquidity of a market. When executing orders in the market, it is important to avoid impacting the market and, consequently, the prices. Usually, the market impact causes adverse price movements implying in an implicit cost of the execution process. Clearly, the intraday trading volume is important when developing execution strategies. In the literature, the intraday trading volume for equities has been reported to possess an intraday U-shaped pattern. In this paper, we investigate for the first time the statistical factors behind the intraday trading volume using non-negative matrix factorization (NNMF). The obtained factors are directly applicable to the design of execution strategies. Additionally, we observe that the factors obtained using NNMF are more intuitively interpretable than the factors obtained using principal component analysis. Our empirical conclusion helps to corroborate several other applications of NNMF from the literature where the same behavior is observed. [ABSTRACT FROM AUTHOR]