Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A product recommendation speech generation method and system based on product selling points

A product recommendation and selling point technology, applied in the computer field, can solve problems such as difficult to achieve smooth sentences, and achieve high controllability

Active Publication Date: 2020-08-04
CHENGDU XIAODUO TECH CO LTD
View PDF17 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, as far as the current technology is concerned, it is difficult to achieve the purpose of fluent sentences only by using the language generated by RNN (excluding translation with more special constraints). When generating poems, even if the grammar is not so correct, it can also It makes sense, for example, "Ancient vines, old trees, faint crows" can even have more artistic conception, so the requirements are not so strict, but when applied to real-life dialogues, it will appear inexplicable

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A product recommendation speech generation method and system based on product selling points
  • A product recommendation speech generation method and system based on product selling points
  • A product recommendation speech generation method and system based on product selling points

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] A product recommendation speech generation method based on product selling points, such as figure 1 As shown, it mainly includes the following steps:

[0060] The first step is to learn the knowledge base composed of prod and attr_value from the basic information of the product. The knowledge base is some basic knowledge, which is very helpful for effect optimization.

[0061] In the second step, we extract descriptions from recommended articles and chat records. The main function is to filter the language and only keep the language that is the recommended language as a description. At the same time, we use the grammar knowledge of nlp to make syntactic adjustments to form a description library.

[0062] The third step is to form a selling point mining model, which can extract relevant selling points from a sentence description. It is also possible to extract relevant selling points from the basic information of the product, and at the same time, there is a ranking of ...

Embodiment 2

[0066] This embodiment is optimized on the basis of Embodiment 1, knowledge base mining:

[0067] The purpose of the present invention is actually to extract what the product is and what selling points it has from the product information. It will be better if we can have a systematic understanding of this aspect. Therefore, it is necessary to mine the knowledge base in advance. The knowledge base we want has prod and attr_value. The demand for attr_name is relatively weak. It is only needed when some attr_values ​​are repeated. For example, the attr_name of clothing is "fabric material" and "lining material", and attr_value can be "pure cotton". , it is necessary to add attr_name at this time. Such as figure 2 As shown, in addition to digging out the complete set of prod and attr_value, the mining of the knowledge base also hopes to dig out the relationship between them. For example, the synonym and implication relationship between prod and prod, attr_value and attr_value...

Embodiment 3

[0078] This embodiment is optimized on the basis of Embodiment 1 or 2, and describes library mining:

[0079] Here, our goal is to find out which sentences are really recommended words. Therefore, it can be regarded as a binary classification model. Such as image 3 As shown, the process is as follows:

[0080] a. Speech segmentation module, which is used to divide a long sentence into several paragraphs and process the corpus more locally. It can be simply split according to punctuation, or some algorithms (such as dependency analysis) can be applied to make the split more reasonable.

[0081] b. The speech classification module uses the short sentences obtained above as our set to be judged, and finds out the beautiful speech that has a recommendation function. A batch of data needs to be labeled first, and then judged by deep learning. Here, we use the BiLSTM+MLP model, which can better combine contextual information through BiLSTM. At the same time, we use word2vector...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention discloses a method and system for generating commodity recommendation words based on commodity selling points. The method includes step S300: mining of selling points: extracting selling points from basic information of commodities, and sorting selling points according to importance; step S400 : Take the direct selling point as the benchmark, find the selling point that can be inferred from the product through the implication relationship in the knowledge base, and use it as the implied selling point; the implied selling point is used to filter the description, and the direct selling point is used to retrieve the description; the product and the description are associated through the direct selling point , to randomly generate the recommended words for the product. The present invention uses the selling point as the key point to generate recommended speech, which will not appear to be just empty language and meaningless; the selling point as the center is the point that buyers are more likely to be interested in, and it will be easier to touch buyers and promote sales went smoothly.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to a method and system for generating commodity recommendation words based on commodity selling points. Background technique [0002] With the rapid development of the e-commerce industry, the proportion of online shopping is increasing, and online customer service plays a very important role. Like offline shopping guide work, introducing products is also one of the important tasks of online customer service. When introducing products, sellers send appropriate copywriting through online dialogue to attract users. The quality of recommendation words directly affects the attraction of products to users power, thereby also affecting the purchase rate of the final commodity. According to experience, good speech skills need to introduce the selling points of the products to customers. Different products have different selling points, and the corresponding speech skills can...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06Q30/06G06F16/242
CPCG06F16/243G06Q30/0631
Inventor 段佳旺江岭
Owner CHENGDU XIAODUO TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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