Clinker Free Lime Predictive Model

• Category: Machine Learning
• Enterprise: CEMEX
• Tools: Python, Scikit-learn

Built and evaluated a machine learning model to support cement quality control by predicting clinker free lime levels from production sensor data. Logistic Regression, Support Vector Machines, and Random Forest were tested, with Random Forest achieving the best results. While model performance was not strong enough for production use, the analysis revealed important data quality and dataset limitations, leading to practical recommendations for improving future model reliability.