Spam is one of the common issues in email systems that can disrupt user
comfort and productivity. This study aims to classify emails into two categories,
namely Spam and Non-Spam, using the Decision Tree method based on the Iterative
Dichotomiser 3 (ID3) algorithm. The classification process is carried out by
designing a dataset consisting of a number of labeled emails and five key attributes,
namely Free, Click, Act, Attached, and Details. The main stages in this method
involve calculating the entropy, average entropy, and information gain of each
attribute to determine which attribute is the most effective in splitting the data. The
attribute with the highest information gain is selected as the root node in the
construction of the decision tree. The entire process is conducted manually to gain
a deeper understanding of how the Iterative Dichotomiser 3 (ID3) algorithm works
in classifying data. The final results show that the constructed decision tree model
can accurately group emails based on keyword patterns, and provides a strong
conceptual understanding of decision-making processes in basic data mining
applications.