Abstract The basis of biometric authentication is that each person's physical and behavioural characteristics can be accurately defined. Many authentication techniques were developed over the years. Human gait recognition is one of these techniques. This article explores machine learning techniques for user authentication on HugaDB database which is a human gait data collection for analysis and activity recognition (Chereshnev and Kertesz-Farkas, 2017). The activities recorded in this dataset are walking, running, sitting and standing. The data were collected with devices such as wearable accelerometer and gyroscope. In total, the data describe 18 individuals, thus we considered each individual as a different class. 10 commonly used machine learning algorithms have been implemented over the HugaDB. The proposed system achieved more than 99% in accuracy via IB1, Random Forest and Bayesian Net algorithms.

中文翻译：

---

机器学习算法实现步态识别

摘要 生物特征认证的基础是可以准确定义每个人的身体和行为特征。多年来开发了许多身份验证技术。人类步态识别是这些技术之一。本文探讨了 HugaDB 数据库上用户身份验证的机器学习技术，该数据库是用于分析和活动识别的人类步态数据集合（Chereshnev 和 Kertesz-Farkas，2017 年）。该数据集中记录的活动是步行、跑步、坐着和站立。数据是使用可穿戴加速度计和陀螺仪等设备收集的。数据总共描述了 18 个个体，因此我们将每个个体视为一个不同的类。HugaDB 已经实现了 10 种常用的机器学习算法。