Analysis of Various Machine Learning Algorithms for Cast Aluminium Alloy to Estimate Fatigue Strength

• Published in: Journal of the Institution of Engineers (India): Series D Vol. 104; no. 1; pp. 61 - 70
• Main Authors: Utpat, Vedant Shrikant, Kulkarni, Swanand Gajanan
• Format: Journal Article
• Language: English
• Published: New Delhi Springer India 01.06.2023; Springer Nature B.V
• Subjects: Algorithms; Alloys; Aluminum base alloys; Artificial neural networks; Datasets; Decision making; Decision trees; Engineering; Fatigue strength; Machine learning; Materials fatigue; Mechanical properties; Metal fatigue; Neural networks; Original Contribution; Programming languages; Python; Regression analysis; Regression models; Root-mean-square errors; Support vector machines; Random forest; Support vector machine; Artificial neural network; Linear regression; Machine learning
• ISSN: 2250-2122; 2250-2130
• DOI: 10.1007/s40033-022-00381-7

In the present work, machine learning algorithms were used to estimate the fatigue strength of cast aluminium alloy. The dataset of 39 alloys of cast aluminium alloys was taken for developing the models. Four machine learning algorithms Linear Regression, Support Vector Machine, Artificial Neural Network and Random Forest was taken for model development. These models were developed by using Python programming language and then evaluation is made by using root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and Coefficient of Correlation ( R ).With each model, fatigue strength was predicted. It has been found that MAPE has the least value for ANN which indicates that, ANN is best suitable model for prediction of fatigue strength. The ANN model is validated with available results which are close agreement with predicted values. Along with ANN, Linear Regression model shows better results after evaluating by RMSE and R evaluation techniques.