The Application Research Of Chinese Word Segmentation Algorithm In GIS

With the advancement of technology and the unceasing development of network technology, geographical information system has applied in every aspect and penetrated into people's daily life. Therefore, providing a Natural Language based Query Interface of geographical information system to users in different stratum has broad prospect of development. The key work of the Interface of geographical information system is translates the natural language to machine language which can be read by computer. So, the primary task is Natural Language Query sentence segmentation and part-of-speech filter, which are the emphases of this article.Firstly, the data logic is divided into common dictionary, space objects dictionary and spatial relationships dictionary, and the various part-of-speech in dictionary is summarized and represented by char. Then, a judgment markup is added in the data structure of Trie Index tree based segmentation dictionary for the subsequent Segmentation Algorithm. The waste of space of Trie Index tree is decreased by high-frequency prefix, while the depth of the tree is also decreased. Secondly, according to the principle of "long words priority" of Reverse-order Directional Maximum Matching, Word Segmentation Method of selecting matching initial word is proposed based on "the longest word possibility", which also deal with the situation of long word cover short word. The time complexity of this algorithm is O(S), which is linear. At last, more parts of speech of data and description problems are solved by shallow part of speech screening algorithm, from which description information which is the unique part of speech of the words in query sentence is obtained. The space complexity is decreased by chain stack structure, and the accuracy of this algorithm is 90%.Natural language query sentence segment and part of speech screening system applied in geographical information system has been designed and implemented. Experimental results show that this system is balanced on velocity and correctness which is effectively enhanced.