| Recently, the discussion about cars’ quality has emerged in the network. According toa survey, there are more than60%customers will get the evaluation information aboutcars from the internet before they purchase a car and they are also regard theseinformation more reliable than the recommends from their friends. Therefore, monitoringthe related IWOM of target cars is very important to the automotive industry. Among themassive network reputation about cars, the negative reputation is more likely to impact theenterprise seriously. Therefore, it is full of significance to construct a specialized,automated negative IWOM identification system to help auto enterprises to analysis thepotential risk exist in network.In this thesis, we design and build a negative IWOM identification system used forautomotive industry,based on the analysis of text recognition technology. The main workof this thesis include the following three parts:(1) Analyzed the existing textcategorization technology and realized the identification of car reputation by trainingSVM package.(2) Analyzed the existing emotional polarity discrimination technology andrealized the identification of negative reputation by using the emotional weight calculatealgorithm.(3) Designed and implemented a negative IWOM identification system used forautomotive industry by combine the text categorization technology, the emotionanalysis technology and the blacklist filter technology.In the end of this thesis, we finished the testing of the system. The testing resultshowed that the negative IWOM identification system used for automotive industry canidentified the target data effectively. Now the system has been used in practice. |
PDF Full Text Request