Core demands
Users cannot quickly and comprehensively understand the experience of tourists when browsing comments. When the homogeneity of a product's comments is serious, it is particularly important to guide users to quickly screen and view comments and fully display the real comments to the greatest extent.
When tourists fill in comments after traveling, they can't remember what to write, lack prompt information, and poor guidance.
To solve these problems, donkey mothers need to interpret all comments, perform operations such as opinion extraction and clustering, and then apply them to experience comment analysis, which requires a lot of technical manpower and material resources, and cannot guarantee the standardization and quality of services.
Solution
1. Through access to Baidu's short text similarity technology, Mommy Donkey Travel Network aggregates similar comments sent by users to avoid serious homogenization of comments on a product, so that buyers can learn about multi-dimensional product comments in a short time and save time in viewing comments.
2. The technology of accessing comment opinion extraction on Mommy Donkey's travel network: extract comment tags from all comments published by users, and mark the polarity of tag opinions (positive/negative). Let each commented product on the platform automatically generate a unique positive or negative comment tag.
For buyers: we can learn about the different word-of-mouth of the product from these labels to help users make consumption decisions.
For new comment users: when filling in comments, you can use the existing comment tags of the product as a reminder to comment on a product more comprehensively from multiple dimensions.
For merchants: we can analyze and monitor the comments of our products according to the labels with opinion polarity that have been extracted, so as to provide reference for subsequent product quality management.
