Automating water quality analysis using ML and auto ML techniques.

Area of Science:

• Environmental Science
• Computer Science
• Data Science

Background:

• Water bodies are increasingly contaminated by industrial and municipal waste.
• Traditional water quality assessment methods are time-consuming and require manual inspection.
• Artificial Intelligence (AI) presents a promising avenue for automating and improving water quality analysis.

Purpose of the Study:

• To compare the performance of Automatic Machine Learning (AutoML) systems against traditional expert systems for water quality assessment.
• To evaluate the effectiveness of AutoML in determining Water Quality Index (WQI) and Water Quality Class.
• To identify the strengths and weaknesses of AutoML in the context of water quality analysis.

Main Methods:

• Utilized Machine Learning (ML) algorithms to build and compare an AutoML system with a custom-built expert system.
• Focused on evaluating the accuracy of both approaches in classifying water quality.
• Assessed performance on both binary and multi-class water quality datasets.

Main Results:

• AutoML and TPOT demonstrated a 1.4% higher accuracy than conventional ML techniques for binary class water data.
• For multi-class water data, AutoML achieved 0.5% higher accuracy, and TPOT achieved 0.6% higher accuracy compared to conventional ML.
• These findings indicate AutoML's potential to enhance the efficiency and accuracy of water quality assessment.

Conclusions:

• AutoML systems show a measurable improvement in accuracy over traditional ML techniques for water quality assessment.
• AutoML offers a more efficient and potentially more accurate alternative to manual inspection and conventional methods.
• Further research into AutoML is warranted to fully leverage its capabilities in environmental monitoring and water resource management.