Using Data compression as a measure of predictability in stochastic stock price data

In this paper we will use a compression algorithm to study the information contained in a stochastic and noisy data. In particular we will use wavelets to compress the random stock price data and study whether the compressibility is a good indicator of the predictability of a price pattern We first apply the compression to data generated from a random walk. We then use this as the base to compare against other data. We will then apply the compression to a regular data set (such sine function). We will also add some noise to a regular function and try and compress that too. Finally we apply this to stock data to see if it reveals a pattern or not.