Weka: Categories of classifiers
Weka is a tool that helps to analyze data sets by using classifiers. A classifier is a machine learning algorithm that learns from the given data and predicts the category of new data. Basically, a classifier studies the data set and learns the relationships within it.
Get Olga Rau’s stories in your inbox
Join Medium for free to get updates from this writer.
For example, we have data about the salary of different employees in the company:
- Age
- Title
- Salary
The classifier can help quickly determine whether a new employee’s salary has been set appropriately by comparing it with historical data, such as the salaries and job titles of previous employees.
Categories of classifiers
Bayes
Uses Bayes’ theorem to predict the most likely class. Example: Naïve Bayes
Functions
It uses equations and functions to classify data. Example: Logistic Regression, SMO.
Lazy
It waits for the new instance instead of building a model during training. Once the instance is there, it compares it to similar examples. Example: IBk (K-Nearest Neighbor).
Meta
It combines multiple classifiers to improve accuracy, so it uses other classifiers as well. Example: AdaBoost
Misc
Just like miscellaneous, a collection of classifiers that do not fit into a specific category. Example: InputMappedClassifier, VFI
Rules
Set of if-then rules. Example: PART
Trees
It repeatedly splits the data and builds a decision tree based on the most informative attributes. Examples: J48, Random Forest
