Search

Answer extraction is an important part of the field of automatic question answering (QA). The traditional answer extraction method relies heavily on contextual semantics, which makes the extraction process consume a lot of time and manpower. Aiming at the question of answer extraction, a method based on the bidirectional long-term memory network conditional random field (Bi-LSTM-CRF) is proposed. This question is extracted in parts from the answer segment. Extract the entity that contains the answer, and then extract the final answer from it. On a valid data set, the accuracy of the experimental results can reach more than 0.6050.

With the development of social media, the development of large-scale social media has become the main platform for public opinion today. Visual analysis of public opinion for keywords under specific conditions uses data mining technology, data cleaning technology and data visualization technology. Through keyword reviews and content mining under specific conditions, discrete data is denoised, cleaned, keyword events and geographic information of interest are extracted. Through the visual expression of the geographical distribution of the focus area, the tag cloud and the geographical map are combined to obtain the public opinion dynamics of keyword events, which opens up ideas for the visualization of public opinion analysis of keywords under specific conditions.

Text to SQL technology is to convert natural language problems into corresponding SQL statements to help users query massive data from the database. This paper introduces the construction of text to SQL data set, the analysis of various technologies of the task model and the application fields based on the technology. It is found that the current model still has a lot of room for improvement for wikisql, spider and other large-scale annotation data sets, while there is little research on Chinese text to SQL technology. Finally, it summarizes some problems that should be paid attention to in future research.

Due to the existing single machine translation model for named entity markup is too rough and not detailed to meet the requirements. This paper proposes an integrated analysis of lexical and syntactic integration of named entity information to improve the performance of Tibetan-Chinese machine translation. Experiments show that the performance of the model is better than that of the model based on independent integration and the model based on independent grammar. In the stage of syntactic analysis, the combination of named entity information and comprehensive information can improve the performance of syntactic analysis.

One potential solution to the problem of information overload is the recommendation system, and the main task of recommendation system based on user portrait is to combine user and product, by mapping user portrait and product portrait, combining association rule mining and reordering algorithm, a personalized product recommendation list is obtained. Not only can help users find valuable information for themselves, but also can make the information displayed in front of interested users, so as to achieve a win-win situation between information consumers and information producers.