In making machine learning models computer generated imagery ai, of course, data is needed. A collection of data used in machine learning is called a dataset, which is then divided/split into a training dataset and a test dataset. The machine learning datasets are used to create/train machine learning models, while the test dataset is used to test the performance/accuracy of the trained/trained models. The technique or approach used to build the model is called an algorithm such as Decision Tree, K-NN, Linear Regression, Random Forest, etc. and the output or result of the process of training the algorithm with a dataset is called the model.
Generally, the dataset is presented in the form of a table consisting of rows and columns. The Column section is the feature or variable of the data being analyzed, while the row section is the data point/observation/example. The thing that is the target of the prediction or the thing that will be predicted in machine learning is called the label/class/target. In statistics/mathematics, this label/class/target is called a dependent variable, and a feature is an independent variable.
A machine learning model that can predict whether a customer will purchase after visiting several e-commerce pages. The target is 1 if the customer makes a purchase and 0 if there is no purchase. Here, the first 10 rows of the dataset are used, Column ProductRelated, BounceRates, ExitRates, and Weekend as predictor variables or features. So, machine learning is divided into 2 types, namely supervised and unsupervised learning. If the label/class of the dataset is known, it is categorized as supervised learning, and if the Label is not known, it is categorized as unsupervised learning.
Recognizing email as spam or not spam is classified as supervised learning because we process datasets containing data points that have been labeled “spam” and “not spam”. Meanwhile, if we want to group customers into several segments based on variables such as income, age, hobbies, or type of work, then it is classified as unsupervised learning.