Text Classification Using Character Frequency Patterns for Spam Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for detecting unsolicited text-based content, particularly in short messages, struggle with low accuracy due to evasion techniques and lack of contextual information, leading to false positives and difficulty in distinguishing between legitimate and unsolicited content.
Innovation Solution
A method involving character category sequencing and frequency distribution analysis is used to classify text-based content, where character sequences are categorized and their frequency distributions are calculated, reducing input dimensionality and computational complexity while maintaining accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If character sequences are used as input for classification, then classification accuracy is maintained, but input dimensionality and computational complexity increase
Solution Approach 1:
The patent transforms the input parameters from raw character sequences to frequency distribution vectors. This parameter transformation reduces the dimensionality of input data while preserving the essential classification information, thereby maintaining accuracy while reducing computational complexity.
Solution Approach 2:
The patent extracts the frequency distribution characteristics from the full character sequences. By taking out only the essential frequency information rather than processing complete sequences, the system reduces input dimensionality while retaining the key features needed for accurate classification.
2Measurement precision
If frequency distribution of character sequences is calculated, then classification accuracy is maintained, but memory utilization increases
Solution Approach 1:
The patent changes the representation parameters from complete character sequences to compressed frequency distribution vectors. This parameter change significantly reduces the memory required to store and process classification data while maintaining the essential information needed for accurate classification decisions.
3Reliability
If traditional detection methods are used for unsolicited content, then detection capability is provided, but false positives occur due to evasion techniques
Solution Approach 1:
The patent transforms the detection approach by changing from analyzing raw text content to analyzing frequency distribution patterns of character sequences. This parameter transformation makes the detection more robust against evasion techniques that manipulate specific characters or words, as the frequency distribution captures overall patterns that are harder to manipulate without detection.
Data Source
AI summary
Approaches to classifying text-based content are described herein. For example, a classification system performs operations that include receiving text-based content comprising a plurality of characters, generating a plurality of character category sequences using the plurality of characters and based on a plurality of predefined character categories, calculating a frequency distribution of the plurality of character category sequences, and classifying the text-based content based on the calculated frequency distribution. The classifying uses a machine learning model that has been trained using a plurality of examples of text-based content. Responsive to the classification, the system can take appropriate actions. For example, responsive to classifying the text-based content as unsolicited, the system can restrict distribution of the text-based content or generate an alert for the text-based content.