15 Oct 2026 Data 2 min read
Finding the goal in the text
A short introduction to the Indonesian SDGs classification project and its Multinomial Naive Bayes approach.
A sample for this website previewPreview article. This sample was drafted from the public project description to review the reading layout. It has not been published by Thoriq.
The Sustainable Development Goals cover 17 areas of global development. A piece of text can discuss one of those areas without naming the goal directly. Classifying that text is one way to make a collection of documents easier to explore.
The Indonesian SDGs classification project looks at this problem for text in Bahasa Indonesia. The public repository documents a training experiment using Multinomial Naive Bayes.
Start with a clear question
The project asks a concrete question: which of the 17 SDGs does an Indonesian text relate to? That question sets the scope of the experiment. It also makes the task understandable before looking at the implementation.
Make the work inspectable
The original project includes a Jupyter notebook. Keeping training work in a notebook lets a reader follow the experiment alongside the code.
For a text classification project, a useful review starts with the input data, the labels, and the way the model is evaluated. The model name is only one part of the picture.
Keep the next question open
This is a project introduction, rather than a claim about model performance. The repository is the place to inspect the implementation and continue exploring the work.