Keyword Stuffing Detection With Block Segmentation and Boundary Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information retrieval systems are susceptible to keyword stuffing, which degrades search accuracy due to the influence of keyword packing on relevance calculations, particularly in retrieval-augmented generation (RAG) systems.
Innovation Solution
A computer-implemented method using a deep neural network (DNN) to allocate character strings into blocks, predict the probability and center/length of keyword stuffing segments, and correct boundaries using local information for improved detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If keyword stuffing detection is performed using traditional methods, then the system can identify keyword stuffing patterns, but the detection accuracy degrades when dealing with long keyword segments and multi-language content
Solution Approach 1:
The patent divides the input character string into multiple blocks of predetermined length, where each block contains a subset of characters. This segmentation allows the detector to process long keyword segments in manageable units, improving detection accuracy for extended keyword stuffing patterns while maintaining the ability to handle multi-language content through localized analysis of each block.
2Device complexity
If the detector processes the entire character string at once, then the system structure remains simple, but the detection capability for long keyword segments is insufficient
Solution Approach 1:
The detector is designed to process character strings in segmented blocks rather than as a single large unit. Each block is independently analyzed by the detector, which maintains structural simplicity while enabling effective detection of long keyword segments through sequential processing of multiple blocks.
3Productivity
If boundary correction is not applied, then the detection process remains fast, but the precision of keyword segment boundaries is inaccurate
Solution Approach 1:
The system performs preliminary detection to identify potential keyword segments, then applies boundary correction as a subsequent refinement step. This two-stage approach maintains initial detection speed while improving boundary precision through localized correction of identified segments.
Solution Approach 2:
The boundary correction mechanism uses feedback from the detection results to refine the identified keyword segment boundaries. By analyzing the detected segments and applying correction based on this feedback, the system improves boundary precision while maintaining overall detection efficiency.
Data Source
AI summary
A non-transitory computer-readable recording medium having stored therein a prediction program that causes a computer to execute a process including allocating an input first character string to a block that satisfies a predetermined condition, predicting, by using a feature amount of each character of a second character string in each block and a detector configured to detect keyword stuffing, a probability that keyword stuffing is present in the second character string, and predicting a center and a length of a keyword segment in the second character string when the probability is a predetermined threshold or more.