PSI - Issue 39

Nabam Teyi et al. / Procedia Structural Integrity 39 (2022) 608–623 Author name / Structural Integrity Procedia 00 (2019) 000–000

618

11

algorithms are applied to a specific data set. ML models can evaluate future unknown data and make predictions or insights, whereas algorithms are simply general approaches to solving an objective. By creating a workspace and partitioning the data into train and test sets, the model is auto-trained. After that, the model is run locally with customized settings based on certain optimization metric. Autogeneration and tuning settings are defined as constraints. As a result, the automatic model is honed. The model results are now examined, and the best model is found. Finally, the most accurate model is put to the test and the desired ML model is obtained (Fig. 10.).

Fig. 10. A machine learning representation.

DTs are a non-parametric supervised learning method for classification and regression (Fig. 11.). Simple decision rules are learned from data properties to build a model that predicts a target variable’s value. A decision tree is a diagram or chart that presents statistical probabilities or helps make decisions. Each branch of the decision tree represents a possible outcome or reaction. These are the outcomes of a given decision path. A RF is a machine learning technique for classifying and predicting outcomes (Fig. 12.). It employs ensemble learning, a technique for resolving complex problems by integrating many classifiers. A RF algorithm is made up of many DTs. The SVM is a linear model for solving classification and regression problems which can tackle both linear and nonlinear problems (Fig. 13.). SVM is a fundamental concept: By drawing a line or hyperplane through the data, the procedure separates it into classes.

Fig. 11. A decision tree representation.