Abstract: The sequence repeating pattern recognition and extraction algorithms have a wide range of practical applications in the fields of data mining,pattern recognition,data compression and bioinformatics．This paper proposed a new algorithm RPT for computing all NE repeats with constraints based on QSA array．The characteristics of the NE repeats are fully considered in the algorithm design procedure for creating a statistical relation between the features and the repeats detection results．Algorithm constraints include minimum period p_min and maximum gap g_max which are used to filter the qualified NE repeats,and algorithm can output all occurrences of the NE repeats in ascending order．Compared with the existing algorithm based on suffix index,the space efficiency of the algorithm is improved．The experiments on classified data sets show that the algorithm RPT is effective for recognizing and extracting the NE repeats on biological sequences,in particular DNA sequences,as well as Uyghur Web texts．

Key words: Repeats,Data mining,Statistical characteristics,Constraints,Bioinformatics,Uyghur Web texts