Prismatic Softwares Pvt. Ltd.

Viscosity Prediction Using Machine Learning

Problem Statement

The goal of this project is to predict the Calculated-IV (Intrinsic Viscosity) of a compound using multiple gel quantity measurements and chemical composition features. The project aims to automate the viscosity prediction process, minimizing the reliance on extensive manual lab testing.

Key Objectives

Automate viscosity prediction process
Optimize data handling and analysis
Improve prediction accuracy

Accuracy

98.5%

Accuracy

98.5%

Integrated Strategy

Data Collection

Assembled and validated chemical composition data for comprehensive analysis

Feature Engineering

Developed relevant features through exploratory data analysis

Model Development

Implemented machine learning models with high accuracy

Tools & Technologies

Python

Pandas

NumPy

Matplotlib

Seaborn

Scikit-learn

Jupyter Notebook

Conclusion

This project successfully transformed raw chemical data into predictive insights using machine learning techniques. The final system supports rapid viscosity estimation, reduces the need for repetitive lab work, and enhances consistency in quality control.

Efficiency

Reduced manual testing time by 75%

Accuracy

98% prediction accuracy rate

Scalability

Handles 1000+ samples/day