Fitting machine learning models

Overview

People use Machine Learning to solve different types of problems. One of its most common applications is prediction. Given some independent observations (usually represented as rows in a table) the task is to predict target values for future observations. To make this prediction a model is trained on seen data (training set). Such data consists of a target variable and predictors, sometimes called features. Features are basically columns used to predict target values. Fitting (training) a model means finding relationship between features and targets. A dataset used for predictions should contain features same as in the training dataset.

How to fit a model

Previously fitting a model required using a programming language like Python or R. Today there are a lot of web-based tools that provide similar functionality using only a graphical user interface. The only problem with those tools is they require data to be transferred and stored somewhere in the cloud. In most cases that means extra registration steps, fees and security issues. With StatSim.Fit you can fit machine learning models online using your machine's resources. That means data is processed locally, no need for registration, and the tool is free.

To fit a model:

1. Select a training dataset in the CSV format
2. Select a dataset for prediction
3. Choose a target variable (the variable you want to predict)
4. Select and customize a model
5. Click Run

That's all!