Method for detecting and filtering unsolicited emails and messages such as spam, phishing attempts and malicious content, involves classifying message as either spam or non-spam using machine learning model, and outputting classification result indicating whether message is spam or non-spam
	DAS S; BISWAS T; BERA T; RAKSHIT S; NAMTIRTHA A; ACHARJEE S; RAKSHIT P
	2025-04-07
专利权人	JIS COLLEGE ENG (JISC-Non-standard)
申请日期	2025-04-07
专利号	IN202531033721-A
成果简介	NOVELTY - The method involves receiving a message in a communication platform. A pre-processing pipeline is applied to a message text. The message text is converted into lowercase. Noise elements are removed. The message text is tokenized. Features are extracted from the pre-processed text using Count Vectorizer to generate a w ord frequency representation and Long Short-Term Memory (LSTM) embeddings to capture contextual patterns. The message is classified as either spam or non-spam using a machine learning model selected from a group consisting of Random Forest and Support Vector Machine (SVM), where the machine learning model is a Random Forest classifier, where the machine learning model is a Support Vector Machine (SVM) classifier. A classification result indicating whether the message is spam or non-spam is output. USE - Method for detecting and filtering unsolicited emails and messages Uses include but are not limited to spam, phishing attempts, malicious content and deceptive advertisements. ADVANTAGE - The method enables leveraging the machine learning and natural language processing techniques to provide an intelligent scalable solution that ensures security and quality of digital communications, while reducing burden on users and platform providers. The method enables offering adaptive monitoring and continuous retraining to handle evolving spam tactics. The method enables providing scalable, efficient and secure solution for modern spam detection in digital communication. The method enables increasing security, efficiency and quality of digital interactions across various communication channels. The method enables achieving an impressive accuracy rate of 97.5%, thus significantly reducing false positives, and hence ensuring that legitimate messages are not mistakenly flagged as spam. The method enables ensuring that users benefit from enhanced security and improved communication quality without the need for technical expertise. DETAILED DESCRIPTION - An INDEPENDENT CLAIM is included for a system for detecting and filtering unsolicited emails and messages. DESCRIPTION OF DRAWING(S) - The drawing shows a flow diagram illustrating a method for detecting and filtering unsolicited emails and messages.
IPC 分类号	G06F-040/00 ; G06N-020/00 ; G06N-003/02 ; H04L-051/21 ; H04L-051/212
国家	印度
专业领域	信息技术
语种	英语
成果类型	专利
文献类型	科技成果
条目标识符	http://119.78.100.226:8889/handle/3KE4DYBR/13195
专题	中国科学院新疆生态与地理研究所
作者单位	JIS COLLEGE ENG (JISC-Non-standard)
推荐引用方式 GB/T 7714	DAS S,BISWAS T,BERA T,et al. Method for detecting and filtering unsolicited emails and messages such as spam, phishing attempts and malicious content, involves classifying message as either spam or non-spam using machine learning model, and outputting classification result indicating whether message is spam or non-spam. IN202531033721-A[P]. 2025.

条目包含的文件
条目无相关文件。

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。