Method, system and device for extracting composable multi-layer text features

The invention discloses a method for extracting characteristics of a multi-layer text capable of being arranged, which belongs to the technical field of text mining in data mining, and can mine potential relationships, models and trends from the multi-layer text capable of being arranged by standardizing an atomic algorithm, arranging the atomic algorithm and executing the atomic algorithm. The accuracy, the extraction efficiency, the data extraction direction and the accuracy of text feature extraction are improved; obtaining a standard algorithm library by using different text extraction algorithms; the method comprises the following steps: obtaining a text extraction event by newly adding a text extraction event and configuring text extraction algorithms needing to be executed in sequence; running the obtained text extraction event to perform text feature extraction on the corpus to obtain an output result; and judging the obtained output result, if the output result can be accepted, completing the extract