Design And Inplementation On Commenting System For Video On Demand

With the rapid development of the internet,the Network for video and audio system is widely used in the entertainment industry,greatly enriching people's leisure time.A large number of illegal message was spread and caused many negative effects with followed.In the hot online video industry,users are often willing to express their emotion when watching the video,but there is some illegal and harmful information flooding.Therefore,it is important and meaningful to provide a healthy and harmonious network environment.The filtering of illegal words among users interaction is an important means to protect the green network environment.Aiming at this problem,commenting system for video on demand have been designed and implemented.The main contents are as follows:1.Based on conditional random field(CRF)model,the thesis implement the process of comments segmentation.The conditional random field model is constructed by reading the CRF model file,and using the characteristic function of the dictionary prefix tree built by double array storage CRF model.It achieve an efficient method for the identification of unlisted words.2.The function of bad information filtering is realized,which combines the basis of finite state automaton algorithm with the characteristics of the bad information filtering.The dictionary prefix tree is constructed with a customize data structure by reading the adverse information lexicon,which can be used to implement the traversal of text.In addition,according to the lexical frequency of harmful information,the tree structure can be modified for improving filtering rate.3.For many models of the client,comment system achieves automatic adaption mechanism.Depending on the type of client,system selects the WebSocket protocol or the long connection mechanism based on the stream to connect and push messages.Server achieve connection pool to avoid frequent connection for sending message process which can reduce server resource consumption.The thesis studies commenting system for video from above three aspects and modifies the original commenting system for HUAWEI video system.The new commenting system improves the bad information filtering efficiency,and implements the bad word filtering commenting system of short text information.In the end,the effectiveness of the system is demonstrated through the experiment.