3  Your task

Your task is to build a simple statistical model to predict transport mode, using the features we discussed in Section 2.3.1. The preprocessed data with features (speed, acceleration, sinuosity) has been provided to you. Focus on the following approaches:

3.1 Aggregate features per segment

Aggregate the point-level features to segment-level summaries. You should:

• Use aggregation functions (mean, max, min, or others you find useful)
• Select which features to include
• Create the training and testing datasets

3.2 Model Building

Use two simple approaches to predict transport mode. And don’t spend too much time on hypterparameter tuning.

3.3 Model Evaluation

• Evaluate the models using appropriate metrics (accuracy, F1-score, etc.)
• Examine confusion matrices to understand which transportation modes are most difficult to classify

4 Submission

Submit the following via Moodle:

A single functioning, annotated computational notebook (e.g. .ipynb, .qmd, or .rmd) structured as a scientific report. In addition, render your computatational notebook to pdf. Submit a zip-file of both the pdf and the notebook. These should include the following chapters:

• Introduction: Introduction to the data and the task
• Material and Methods: Computing environment, data description, statistical methods, and commented R code for feature aggregation, model training, and evaluation (as specified in Chapter 3)
• Results: Text, figures, tables presenting model performance metrics, confusion matrix, and plain language summary of findings (no visible code)
• Discussion: Markdown text only (no figures or tables) discussing interpretation, limitations, and potential improvements