Filtering financial time series by least squares

Modeling of financial time series with artificial intelligence is difficult due to the random nature of the data. The moving average filter is a common and simple form of filter to reduce this noise. There are several of these noise reduction methods used throughout the financial security trading community. The major issue with these filters is the lag between the filtered data and the noisy data. This lag only increases as more noise reduction is desired. In the present marketplace, where investors are competing for quality and timely information, lag can be a hindrance. This paper proposes a new moving average filter derived with the aim of maximizing the level of noise reduction at each delay. Comparison between five different methods has been done and experiment results have shown that our method is a superior noise reducer to the alternatives over short and middle range lag periods.