There are certain programs which each have certain features. There is also a feature vector which is the Union of all the features. I have modeled the vector of features as following.
1 - the feature is present in the feature vector
0 - the feature is absent in the feature vector.
I have to find a relation between the following variants
10001110
10110101
10010111
10101100
10011101
Here the relation in all the programs is that the 0,1 and 5 positions do not change i.e., the features they represent, tend to continue. Given a large data set as this are there any Machine Learning/Data Mining approaches that can find the relation between them.
You might use affinity analysis, e.g., as used in market basket analysis.
Suppose each vector of yours is a shopping bag checked out by a customer, and each column is a product. Then market-basket analysis will find stuff like:
Buying milk + eggs => buying butter
Most by bread
The application to your problem seems natural.