| With the development of social economy and the improvement of people’s living standards,cars have become an indispensable means of transportation for people.The demand for cars is increasing,but there are many information about cars(such as car brands,models and related prices),Users cannot accurately find the vehicle that suits them.At present,many scholars and researchers have invested in the research,but in the current era of big data,people still cannot use traditional search engines to accurately retrieve the results they want.When users use traditional search engines to search for problems,it returns to the user’s webpage with numerous vehiclerelated information.Among them,it contains information that is not beneficial to users,and the user needs to perform secondary screening on the results,wasting user time.In order to allow users to quickly obtain information and improve user experience,this paper designs and implements a smart question and answer system based on TensorFlow,make up for the shortcomings of traditional search engine which provides accurate answers to user questions and reduces user waiting time.The paper introduces the background and research status on the subject,summarizes the current problems of the question answering system,and combines the development status of similar systems at home and abroad,based on solving the actual needs of users about automotive problems,and combs the business process of the question answering system.The author use UML language to build the system function module,use the TensorFlow deep learning framework to design and implement a deep learning based automotive intelligent question answering system.The main research work of this paper : first uses crawler technology to grab the corresponding data set from the larger car portal on the Internet,and preprocesses it to finally get a data set with a uniform data format;after using the Word2 Vec model to convert the word vector to make the expression of the Chinese word vector more Accurate;then select the TensorFlow framework to build the Seq2Seq(Sequence to Sequence)model.Based on this model,use a Bidirectional Long Short-Term Memory model,and introduce an attention model on the Bidirectional Long ShortTerm Memory model,so that the model can provide more accurate according to the focus of the sentence.result.Finally,in the system’s human-computer interaction interface,an app is designed using Android technology.The car intelligent question answering system based on deep learning designed and developed in this paper can synthesize many kinds of information,accurately and quickly provide users with reliable information. |
