In this paper, we introduce an approach to secure IoT devices from unsolicited emails by using certain AI-based features and clustering in real-time. We propose a novel approach that first filters the unwanted emails from the incoming traffic and then classifies them into spam and phishing for Internet of Things (IoTs) based systems in industry 4.0. The AI mechanism collects and analyzes emails to detect multiple features that identify patterns for classification. We divided our incoming data into batches and each batch was classified based on knowledge gained from previous batch's classification. We tested our results with a number of classifiers and results show that our approach gives highly accurate classification.

在本文中，我们介绍了一种通过使用某些基于AI的功能和实时群集来保护IoT设备免受未经请求的电子邮件的攻击的方法。我们提出了一种新颖的方法，该方法首先从传入的流量中过滤掉不需要的电子邮件，然后将它们分类为针对工业4.0中基于物联网（IoT）的系统的垃圾邮件和网络钓鱼。AI机制会收集和分析电子邮件，以检测多种识别分类模式的功能。我们将传入的数据分为多个批次，并且根据从先前批次的分类中获得的知识对每个批次进行分类。我们使用许多分类器测试了我们的结果，结果表明我们的方法可以提供高度准确的分类。