Abstract—

In the present decade, the quality of water has become the major concern, because of the rapid increase of pollution on water resources. This has been a great threat for all living beings in the planet. Hence, there is always a demand on an efficient model for water quality analysis. With that note, this paper develops an enhanced automated machine learning model called EAMLM. Moreover, the proposed model utilized Internet of Things (IoT) based sensors to obtain the quality factors of water such as pH rate, temperature, nitrogen, phosphorous, total hardness and total dissolved solids (TDS). The model integrates the IoT analysis with the operations of machine learning methods to evaluate the real-time data of water samples obtained from local areas. In particular, the EAMLM algorithm is framed with the combined efficiencies of modified ranking based K-nearest neighbors and random forest (RF) model. Further, Raspberry Pi3 is low cost kit embedded for sample testing and the model is simulated and evaluated using WEKA tool. The classification results show that the EAMLM provided better accuracy than other traditional models.

中文翻译：

---

Eamlm：使用实时数据集进行基于物联网的水质分析的增强型自动化机器学习模型

摘要-

近十年来，由于水资源污染迅速增加，水质量已成为人们关注的焦点。这对地球上的所有生物来说都是一个巨大的威胁。因此，始终需要一种有效的水质分析模型。鉴于此，本文开发了一种名为 EAMLM 的增强型自动化机器学习模型。此外，该模型利用基于物联网（IoT）的传感器来获取水的质量因素，例如pH值、温度、氮、磷、总硬度和总溶解固体（TDS）。该模型将物联网分析与机器学习方法的操作相结合，以评估从当地获取的水样的实时数据。特别是，EAMLM 算法结合了基于 K 最近邻的改进排序和随机森林 (RF) 模型的组合效率。此外，Raspberry Pi3 是嵌入用于样本测试的低成本套件，并使用 WEKA 工具对模型进行模拟和评估。分类结果表明，EALMM 比其他传统模型提供了更好的准确性。