E-commerce Review Spam Filtering via User ID Library

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for filtering advertisement and spam reviews in electronic commerce are inaccurate due to reliance on manpower for dictionary establishment and require large amounts of labeled data for machine learning, leading to low identification rates for irregular and unpredictable spam reviews.

Innovation Solution

A method and system that acquire predetermined advertisement spam samples, establish a user identification library, and classify new reviews based on user identification, utilizing a text classifier and spam review text analysis to determine advertisement spam reviews, improving accuracy and recall rates.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a manually established dictionary method is used to identify advertisement reviews, then the identification process can be simple, but the accuracy and recall rate are low due to subjective keyword extraction and incomplete coverage

Engineering Contradiction:

Improvesimplicity of identification processVSAvoidaccuracy and recall rate of identification

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system enables automatic identification of advertisement spam reviews through machine learning classification, eliminating the need for manual dictionary establishment and keyword extraction. The model learns patterns from labeled data and autonomously classifies new reviews, achieving both high accuracy and operational simplicity.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual process of dictionary creation and keyword matching with an automated machine learning system. The mechanical system of manual observation, extraction, and compilation is substituted by computational algorithms that automatically learn from data and perform classification.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If machine learning classification is used to identify advertisement and spam reviews, then the identification accuracy improves, but large amounts of labeled training data are required which increases system complexity

Engineering Contradiction:

Improveidentification accuracyVSAvoidsystem complexity due to data labeling requirements

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary classification of review texts into advertisement spam and normal categories using a trained machine learning model. This preliminary action enables subsequent targeted processing, such as user identification library updates, based on the classification results.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback mechanisms where classified advertisement spam reviews are used to update the user identification library, which in turn improves future classification accuracy. The continuous learning and updating process creates a positive feedback loop that enhances system performance over time.

Inventive Principle:

Principle #23Feedback

3Productivity

If traditional text classification algorithms are used, then the system can process reviews automatically, but the identification rate for irregular and unpredictable spam reviews remains low

Engineering Contradiction:

Improveautomatic processing capabilityVSAvoididentification rate for irregular spam reviews

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent employs a composite approach combining multiple classification strategies: machine learning text classification, user identification library matching, and keyword-based filtering. This composite system leverages the strengths of each method to achieve high identification rates for diverse spam types, including irregular and unpredictable reviews.

Inventive Principle:

Principle #40Composite materials

Solution Approach 2:

The system adds a new dimension to spam identification by incorporating user identification information alongside text content analysis. By examining the user who posted the review in addition to the review text itself, the system gains another dimension of information for classification, improving detection of irregular spam reviews.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10963912B2Method and system for filtering goods review information

Publication Date: 2021.03.30 BEIJING JINGDONG SHANGKE INFORMATION TECH CO LTD

AI summary

The present invention discloses a method and a system for filtering goods review information. The method comprises: acquiring a plurality of predetermined advertisement spam samples, each advertisement spam sample comprising a review text and a user identification; establishing an advertisement spam user identification library comprising the user identifications of the plurality of advertisement spam samples; and acquiring a new review comprising a user identification and a review text, and determining the new review as an advertisement spam review if the user identification of the new review is included in the advertisement spam user identification library. An advertisement spam review is identified according to a user identification that publishes the review in the present invention. A new method is provided in the technical field of identifying an advertisement spam review for solving the problem that messy spam reviews are difficult to identify.