Research On Online Review Of Homestay Based On Text Mining Technology

In the context of sharing economy,with the booming development of China's tourism industry,homestay industry has developed by leaps and bounds in recent years.The development of information technology and the multifaceted penetration of social platforms,It also makes online booking of accommodation become a mainstream way.Compared with offline hotel booking,online booking attracts more consumers with its advantages of transparent price and rich choices.At the same time,the number of online reviews the state of the exploding,review data contains abundant information,reflects the consumer the most intuitive experience for homestay facility,we can get from comments for home stay facility hold the attitude of consumer,through analysis the data,can find the advantages and existing problems of homestay facility industry,as to optimize the service quality of service providers,continuously meet the needs of customers,establish a good image of the property in the consumers' mind,thus improving customer loyalty,promote the benign development of short online rental industry.Based on this background,this paper takes the comment data of the users on the home stay website as the research object.Firstly,it obtains the comment data through the web crawler technology,and then preprocesses the data such as deduplication and Chinese word segmentation.In this paper,the sentiment dictionary and machine learning method are used to classify the sentiment of comment data,and the comment text is divided into positive and negative categories.Considering that the classification method based on machine learning requires manual annotation of data sets in advance,Therefore,the method of constructing emotion dictionary is firstly used for preliminary classification of text data,and then class labels are obtained,so as to avoid the tedious process of manual labeling.Then,the marked data set is divided into training set and test set according to a certain proportion.After a series of processing,such as data cleaning,word segmentation,word removal and text feature extraction,The classification algorithm is used to train the training set,and then the sentiment classifier is obtained.The test set is used to verify the classification performance of the sentiment classifier.And then based on the data set that's already classified,Using a variety of visualization methods,such as word cloud generation and semantic network construction analytics,Further in-depth characteristic analysis was conducted on different types of review data,Obtaining the experience information and product characteristics information of consumers in the process of living in the homestay.Finally,through the construction of LDA topic model,in-depth semantic mining analysis of comment information is carried out.And according to the results of the analysis to the merchant put forward feasible suggestions.Research results show that the current homestay industry in the consumer reputation is still good,most consumers are positive attitude to homestay,of course,there are some consumers in the homestay experience is not satisfied.In addition,through the analysis also found that,when consumer is choosing a home stay facility for room hardware facilities,geographical location,health status and service attitude is very seriously,the home stay facility basically meet the needs of the consumers in these areas,but needs to be improved in some detail aspects,such as interior old room,air conditioning failure,hot water supply,sound insulation effect is poorer,and the bed is tasted enough cleanliness and comfort,etc.