Design And Implementation Of The Information Processing System Of Safety Accidents Based On Understanding Natural Language

In terms of the occurrence of safety accidents,China has already far exceeded the main developed countries.The malignant development of this kind of situation is not conducive to the stability of people's lives,but also is not conducive to the sustainable development of the country.Therefore,it is urgent to take effective means to manage the safety accident information.But the management and application of the domestic relevant departments' website for safety accident information indicate that the Department did not fully explore the intrinsic value of the data,it also has little contribution to the prevention of safety accidents.In this paper,the design and implementation of safety information processing system based on natural language understanding,the system is responsible for the safety accident information from2002 to 2016 for processing and visualization,Among them,"Processing" refers to the extraction and classification,"visualization" refers to the design and implementation of the system according to the process flow,statistics and analysis of the safety accident information.First of all,Traversing and judging the formation of word segmentation sequences after word segmentation,so as to obtain the information corresponding to the time and place.Natural Language Processing in POS tagging technology can be used to traverse and get all numeral segmentation in segmentation sequence.In the word segmentation sequence,it is through the judgment of the left and right side of the word segmentation to determine whether the word can represent the "injured","death" and so on,In this regard,an incremental search algorithm is proposed to extract the information of safety accidents.However,there are problems in the location of the results of information extraction,such as "missing provinces" and "empty address",in this paper,we identify all the address entities in the word segmentation sequence and then analyze the frequency of the administrative units in each level,and use it as the basis for judging the final address of the information.In this regard,"address complete algorithms" is put forward to be used to extract the key information.Secondly,this paper proposes a "selective extension algorithm",considering the characteristics of safety accident information and the effects of redundant word segmentation and sparse features,this algorithm presents a short text feature vector representation,it is using support vector machine to complete the classification ofsafety accident information.Finally,the web system is designed and implemented,which can provide decision support for the prevention of safety accidents.