Question

What is different in letting machine learning algorithms to learn about numeric data versus categorical data?

Answer 1

Answer:-

However, there are appropriate and not-so appropriate algorithms for particular problem tasks. I mean, every dataset is numerical when you use machine learning. If you have categorical "string" variables, you typically encode them in some way (e.g., one-hot encoding). So, the choice of algorithm highly depends on

- the size of the training data
- the number of features
- the quality of features
- the number of unique class labels
- linear vs. non-linear problems

A lot of it depends on the “shape” of the data. The simpler the trends in the data are, the simpler of a model you can use. And in general, simpler is better, as long as the model is capable of capturing the trends in the data. Here’s a good visual intuition of many common machine learning algorithms.

comparison of a several classifiers in scikit-learn on synthetic datasets. The point of this example is to illustrate the nature of decision boundaries of different classifiers. This should be taken with a grain of salt, as the intuition conveyed by these examples does not necessarily carry over to real datasets.

Particularly in high-dimensional spaces, data can more easily be separated linearly and the simplicity of classifiers such as naive Bayes and linear SVMs might lead to better generalization than is achieved by other classifiers.

The plots show training points in solid colors and testing points semi-transparent. The lower right shows the classification accuracy on the test set.