JASP guides you step-by-step through advanced analyses, from Bayesian stats to predictive modeling, and presents the results in a clear, readable way. It even automatically handles tasks like splitting your dataset into training sets (to build the model) and testing sets (to see how well it performs).

Advantages of JASP over SPSS

JASP was developed at the University of Amsterdam and is now used all over the world. The program supports standard analysis procedures in both their classical (frequentist) and Bayesian form. It offers some major advantages over SPSS, such as state-of-the-art meta-analysis and compatibility with APA-publishing guidelines. One of the motivations for the development of JASP was to make it easier for statistical practitioners to conduct Bayesian analyses. Bayesian statistics offer a flexible approach to analysis, allowing you to incorporate prior knowledge into your models and interpret results in terms of probability. This can be especially helpful when data is limited or uncertain.

Why use JASP for Machine Learning?

The Machine Learning module in JASP allows you to:

• Work with real data using easy point-and-click menus
• Explore classification, regression, and clustering models (see explainer section on Machine Learning)
• Interpret model results without technical jargon

Example: Predicting Customer Churn

Imagine you work for a telecom company and want to predict which customers are likely to cancel their subscriptions. This is known as a classification task. In this example, the Telco Customer Churn dataset is used to train a model that predicts whether a customer will churn (“Yes” or “No”) based on their contract details and demographics.

JASP’s Machine Learning module makes this surprisingly easy. One commonly used method is K-Nearest Neighbors (KNN), which classifies a case based on how similar it is to other cases in the data. Another option is Random Forest, a technique that builds lots of decision trees and combines them to improve accuracy, great for handling complex or messy data.

Step-by-Step: Train a Classification Model in JASP

The JASP team provides great guides to get you started. This example is based on JASP’s tutorial on How to Train a Machine Learning Model in JASP: Classification - JASP - Free and User-Friendly Statistical Software.

1. Open JASP
   Download and launch JASP. It’s free and open-source.
2. Load the dataset
   Go to File > Open > Data Library
   Navigate to 10. Machine Learning, and open the Telco Customer Churn dataset
3. Explore the data (Optional but Recommended)
   Use the Descriptives analysis to check how many customers have churned (“Yes”) vs. stayed (“No”). This gives you a quick feel for your data and reveals class imbalance.
4. Go to Machine Learning module
   Click on the Machine Learning tab in the top menu.
   Select K-Nearest Neighbors Classification (or another classification analysis like Random Forest).
5. Set the target and predictors
   Target: Select Churn (this is what you want to predict)
   Features: Select all variables except customerID.
6. Configure data splitting
   In Data Split Preferences, you can change the percentage holdout test data. This is the percentage of unseen data that the quality of the trained classification model is evaluated with.
7. Set model parameters (optional)
   Under Training Parameters, choose a fixed value for the number of nearest neighbors (e.g. 3).
8. Run the model
   Once you’ve selected your target and features, JASP automatically runs the analysis and shows the results in tables and plots.
9. Evaluate model performance
   Scroll down to view key statistics like accuracy, sensitivity, specificity, and confusion matrix. Use these numbers to assess how well your model predicts customer churn.
10. Save or export your results
    Click File > Save As to save your JASP file, or export your outputs as PDF, HTML, or LaTeX.

🎓 Bonus tip: Try different algorithms (e.g., Random Forest, Logistic Regression) and compare results - all within the same interface.

Try it yourself!

Follow the steps above to build your first classification model and get familiar with JASP. Once you’re comfortable, experiment with other algorithms or try a regression or clustering workflow. Happy analyzing! And don’t forget to check out all the other great blogposts by JASP to play around or maybe even become a pro.

Welcome to the ‘Library Research Hack,’ the Library’s feature designed to take your research to the next level. Whether you’re a student or a researcher, each month you’ll discover a new tool or application that enhances your research efficiency, effectiveness, and fun. From databases that expand your horizons to essential research identifiers and powerful software for data analysis, we explain why these resources are indispensable and guide you on how to get started.

Stay tuned – we'll be back in June with a new hack on Think. Check. Submit.!