Identification Server for Fraudulent Call Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods lack an efficient defense mechanism against fraudulent telephone calls, which often target vulnerable individuals, causing financial and psychological harm, as subscribers typically rely on recognizing the fraud themselves.
Innovation Solution
A method involving an identification server that analyzes call information data using call pattern, voice, and context analysis to identify unwanted calls, employing machine learning and neural networks to detect anomalies and initiate countermeasures such as warnings or call blocking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated identification systems are implemented to detect fraudulent calls, then protection effectiveness is improved, but device complexity increases
Solution Approach 1:
The patent introduces an identification server as an intermediary component between the telecommunications network and the terminal device. This server performs call pattern analysis, voice analysis, and context analysis to identify fraudulent calls, thereby protecting subscribers without requiring complex functionality in their terminals. The intermediary handles the complexity centrally while keeping end-user devices simple.
Solution Approach 2:
The identification system is divided into distinct functional modules: call pattern analysis unit, voice analysis unit, context analysis unit, and probability calculation unit. Each module performs a specific analysis function and contributes to the overall fraud detection. This segmentation allows the complex protection function to be implemented as a coordinated system of simpler, specialized components.
2Measurement precision
If multiple analysis methods are used to improve detection accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system performs three types of analysis (call pattern, voice, and context) and combines their results through probability calculation. While full implementation of all analysis methods provides high detection accuracy, the system can operate with varying degrees of analysis depth depending on the situation, balancing precision requirements with computational resources available.
Solution Approach 2:
The system calculates a probability value that combines results from multiple analysis methods and compares it against a threshold to determine whether a call is fraudulent. This feedback mechanism allows the system to dynamically adjust its detection decisions based on the combined evidence from different analysis sources, improving overall accuracy while managing complexity through structured decision logic.
3Speed
If real-time analysis is performed to protect subscribers immediately, then speed of response is improved, but use of energy increases
Solution Approach 1:
By placing the computationally intensive analysis functions on a centralized identification server rather than on the subscriber's terminal device, the system enables real-time protection without requiring the terminal to consume excessive energy. The server performs the heavy processing while the terminal merely communicates with it, maintaining fast response times while distributing energy consumption to the network infrastructure.
Data Source
Figure 1
AI summary
The present invention relates to techniques for identifying unwanted calls to a call-capable terminal of a subscriber in a communication network, comprising the following steps: • Receiving call information data characterizing the call made by an identification server of the communication network, wherein an algorithm for identifying unwanted calls is implemented on the identification server, the algorithm being programmed to execute at least one of the following identification methods: • a call pattern analysis, wherein the call pattern analysis is configured to compare a typical call pattern known via the terminal with the call information data and to identify an unwanted call by the comparison;and/or ∘ voice analysis, wherein the voice analysis is configured to extract speech as audio signals from the call information data and compare them with known speech profiles, and to identify unwanted calls by comparison; and/or ∘ context analysis, wherein the context analysis is configured to: in a first step, extract speech as audio signals from the call information data; in a second step, convert the speech into text; and in a third step, analyze the text contextually to identify unwanted calls.;