Design And Implementation For Prediction Technique Of Fault Over Set-Top Of IPTV

As compared to the conventional television,Internet Protocol Television(IPTV)supplies the much higher quality of the picture,gives the diversified resources from the internet,and achieves the interactive experience for the audience.However,the IPTV encounters some technique difficulties as a result of the high volume of data,especially,the prediction,localization and fast solving over the fault of the set-top box from IPTV.Therefore,the design and the implementation are studied in this thesis for the prediction technique of the fault over the set-top from the IPTV.The key focus of this thesis is placed on the design of the fast prediction for the fault of the set-up in IPTV systems,where the principle of the machine learning is employed.Also,the corresponding implementation is achieved.As compared to the methods by hands for the fault prediction over the set-up of IPTV,the method proposed in this thesis have the advantages of both the lower cost for analysis and the higher precision.Moreover,the cost of the fault prediction is also reduced.First of all,the mathematical model for the fault is formulated for the set-up of IPTV.Then,the strategy is designed for the data acquisition,which is followed by the design of the save structure.Finally,the design and the implementation is carried out for the fault prediction of the set-up in IPTV systems,which is based on Spark.The property of the statistics is exploited from the data of the operation and maintenance by employing Spark,based on which both the data save and the acquisition are achieved.The design of the feature extraction is completed.The method is established for the fault prediction.The performance test is carried out for the fault prediction in the set-up of IPTV.The precision performance is improved for the fault precision.Therefore,the labor cost is reduced.The service requirement is substantially satisfied for the fault diagnosis and prediction of the set-up in IPTV systems.First,the background of the fault prediction is reviewed for the set-up of IPTV system,where the typical models are summarized for the fault prediction in this system.Then,the model prediction algorithms,feature extraction algorithms and model test algorithms are discussed,which are based on Spark.In what follows,the model is explained in detail for the fault prediction in set-up of IPTV.Finally,the implementation framework is presented for the fault prediction technique in the set-up of IPTV,where the corresponding performances are tested.