A sentiment analysis method and device for automobile word-of-mouth
A technology of sentiment analysis and automobile, applied in the field of sentiment analysis of automobile word-of-mouth, can solve the problems of incomplete sample data, troublesome, large maintenance, etc., achieve the effect of improving classification accuracy, solving memory problems, and avoiding overfitting phenomenon
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0071] see figure 1 , the present invention provides a kind of embodiment one of the emotion analysis method of automobile word-of-mouth, comprising:
[0072] Step 101, obtaining car word-of-mouth data for training and testing from the car platform.
[0073] It should be noted that before analyzing the word-of-mouth of new cars, a hidden Markov model for analyzing word-of-mouth of new cars must be constructed. Therefore, in this application, the data of word-of-mouth of cars used for training and testing is first obtained from the car platform. .
[0074] Step 102 , based on natural language processing, perform entity extraction of car configuration items, emotional words, degree words, and negative words from the car word-of-mouth data, and obtain sample data after judging the emotional polarity of the cut corpus where the car configuration items are located.
[0075] In this embodiment, after the car word-of-mouth data is obtained from the car platform, based on natural la...
Embodiment 2
[0089] see figure 2 , the present invention provides a second embodiment of the sentiment analysis method of automobile word-of-mouth, including:
[0090] Step 201, obtaining car word-of-mouth data for training and testing from the car platform.
[0091] Understandably, if image 3 Shown is a functional block diagram of the car word-of-mouth sentiment analysis in this embodiment.
[0092] Specific semantic analysis such as Figure 4 , Figure 4 Among them, the ordinate is the first-level indicator in the original comment, the abscissa is the first-level indicator after proofreading through the program, and the color code represents the correlation, here multiplied by 100, the sum of each row of indicators is 100, and the sample size is randomly selected 6 10,000 comments, for example, 96.79% of Space has been judged as Space after correcting the corpus, and 3.21% has been judged as Cost-effective, indicating that the user has written the description under the topic of Cos...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com