Currently there is the distribution analysis, word stemming and edit distance implemented. This provides a way of checking similarities in different ways. What we still need to do is to generalize these solutions and making them work together. The problem is that all this is hard coded. Question is how we can improve this situation up to a point where the process of pattern recognition discovers new ways of how to recognize patterns and similarities automatically.
Currently there is the distribution analysis, word stemming and edit distance implemented. This provides a way of checking similarities in different ways. What we still need to do is to generalize these solutions and making them work together. The problem is that all this is hard coded. Question is how we can improve this situation up to a point where the process of pattern recognition discovers new ways of how to recognize patterns and similarities automatically.