I have been trying to adapt prefixspan algorithm for mining labeled sequential patterns and labeled tree patterns by incorporating a minimum confidence threshold. it works fine for for sequential patterns but there is a slight problem with the later. the original paper suggests that an itemset cannot have an item repeated. but when mining syntactic trees, there is a high chance that it will happen. ex. (S(NP(PRP)(PRP)(NN)(.)), E)

can someone help!
[mineLTP] algorithm can be found at http://www.aaai.org/Papers/AAAI/2007/AAAI07-147.pdf