Prediction is a method of detecting a person's behavior toward online buying by evaluating publically available evaluations on the web. Understanding expressive human communication involves a simultaneous examination of speech and gestures since human behavior is communicated through a combination of verbal and nonverbal channels. Machine learning algorithms are utilized in this work to extract evaluations from the net and categorize these into five classes, namely, highly favorable, favorable, neutrality, bad, and strongly negative, for the prediction of human behavior. A person's behavior is analyzed, and the experimental comparison is made to machine learning methodologies. Numerous classifiers are employed on manuscript transcripts in this work to determine the accuracy, exactness, recollection, and f1-score, which are also represented in terms of muddle matrices. In this research studies a new technique is proposed using Ludwig classifier and it has been discovered that deep learning employing the Ludwig classifier achieves near-perfect accuracy about 99.9%. The outcomes are presented to demonstrate the preceding point.

预测是一种通过评估网络上公开可用的评估来检测一个人的在线购买行为的方法。理解富有表现力的人类交流涉及同时检查语音和手势，因为人类行为是通过语言和非语言渠道的组合进行交流的。在这项工作中使用机器学习算法从网络中提取评估并将其分为五类，即高度有利、有利、中立、坏和强烈否定，用于预测人类行为。分析一个人的行为，并与机器学习方法进行实验比较。在这项工作中，手稿抄本上使用了许多分类器来确定准确性、准确性、回忆和 f1 分数，也可以用混淆矩阵来表示。在这项研究中，提出了一种使用 Ludwig 分类器的新技术，并且发现使用 Ludwig 分类器的深度学习可以达到近乎完美的准确率，约为 99.9%。呈现结果是为了证明上述观点。