Research On Heterogeneous Data Based Investors' Interactive Communication System And Its Implementation

The investors' interactive communication System developed by Shenzhen Securities Information Co.,Ltd.(SSIC)provides a direct communication platform for individual investors and listed companies.The paper elaborates the upgrading of the technical architecture of the original system and the issues of full text retrieval and real-time data updating addressed in the process of constructing new generation interactive communication system.The main research and developments are as follows:An optimized model is put forward for the ranking of full text retrieval result according to relevance of communication contents.Solr,a Lucene-based search server,is used in the new interactive communication system,but the data timeliness of securities market is ignored in its standard formula,so the formula is reconstructed in accordance with the position weight and time weight of key words in a text to solve the issues of contents relevance and time relevance,so as to enhance the search accuracy of investors' interactive contents.Moreover,in order to help investors to know the stock market fad and listed company news,a real-time updated optimization model based on investor relations data is put forward in the paper.On one hand,the storm stream processing model along with popular factors statistics is used for real-time updates for association relationship between investors,investors and listed companies,and investors and topics or circles they are involved in.On the other hand,an optimized formula for statistics of heat ranking list is put forward to optimize the ranking result of the list based on the business weight coefficient and type index value.Investor interaction and relations data not only include structured basic data(such as information of listed companies,senior management and individual investors,etc.),but also include unstructured extended data(such as full text of announcement,picture and video,etc.),and different kinds of data have different update frequency.In order to meet the requirements of interactive website for high-efficient storage,high-concurrency reading and writing,high availability and extensibility,a storage technical solution of ‘relational database,documental database and distributed file system' is introduced,including inter-table mapping rules design for historical data import and data synchronism optimization after its formal launch.The combination of the subject research and interaction system for investors application help improve the accuracy of contents relevance in full text retrieval,enhance the statistical efficiency of investor exchange data and make the new interactive communication platform better meet the requirements of institutional regulation users and external individual investors for internal function extension and maintainability through technical architecture optimization in the foreground and storage structure optimization in the background.