Hi Fawce,
If I've coded things correctly, I should be comparing each of the five securities to SPY with:
for j in range(m):
X = ''
Y = ''
for i in range(W_L):
X = X + str(int(coded_prices[i,0]))
Y = Y + str(int(coded_prices[i,j+1]))
N[j] = NCD(X,Y)
In the code snippet above, X is the text string representing the coded volume of SPY over the 30-day sliding window. Y is the text string of the coded volume of one of the five securities (I loop over the five securities, storing the result for each security in N).
The basic idea is to get a sense for which securities trade like SPY and which ones do not. For example, AAPL seems to stay within a narrow range, while other securities spike up, sometimes becoming significantly less similar to SPY in their volume patterns.
I have no idea if this will bear fruit...I'm just tinkering around learning how to do pattern recognition (I think...).
Grant