++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
This folder contains the data used for the paper titled "I Can Tell by Your Eyes! Continuous Gaze-Based Turn-Activity Prediction Reveals Spatial Familiarity," published in LIPIcs, Volume 240, COSIT 2022.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1. the model was trained on the training dataset named "export_feature_train.csv" and tested on the data of each experiment separately, i.e. "export_feature_test.csv" in each of the subfolders
2. the subfolders are named as "id_cond", where id is the route identifier and cond is the trial condition, i.e., familiar (fa) and unfamiliar (un).
3. the columns in the training/testing dataset are as follows:
	3.1. participant_id: unique ID for each participant.
	3.2. condition: condition as a binary measure (familiar and unfamiliar)
	3.3. window_id: the ID of the window for sliding window segmentation
	3.4. route_id: unique ID for each route
	3.5. num_ways: number of road segments at each intersection
	3.6. delta: angular measure of the road segments at each intersection
	3.7. delta_t: deviation of straightness of road sections at each intersection from a T-intersection
	3.8. osm_id: object ID in the Open Street Map data source
	3.9. junction_nb_along_route: Number of the intersection along the route
	3.10. distance_along_route: distance of the current intersection from the beginning of the route
	3.11. TDP: turn decision point with three values (TL: turn left, TR: turn right, NT: do not turn)
	3.12. gps_time_one_hour: gaze time synchronized with the time recorded by GPS
	3.13. accuracy: gps accuracy
	3.14. fixation_frequency: frequency of fixations in each window
	3.15. mean_fixation_duration: average duration of fixations in each window
	3.16. min_fixation_duration: minimum duration of fixations in each window
	3.17. max_fixation_duration: maximum duration of fixations in each window
	3.18. var_fixation_duration: variation in duration of fixations in each window
	3.19. mean_fixation_dispersion: average dispersion of fixations in each window
	3.20. min_fixation_dispersion: minimum dispersion of fixations in each window
	3.21. max_fixation_dispersion: maximum dispersion of fixations in each window
	3.22. var_fixation_dispersion: variation in dispersion of fixations in each window
	3.23. mean_fixation_dispersionX: average dispersion of fixations along the x-axis in each window
	3.24. min_fixation_dispersionX: minimum dispersion of fixations along the x-axis in each window
	3.25. max_fixation_dispersionX: maximum dispersion of fixations along the x-axis in each window
	3.26. var_fixation_dispersionX: variation of dispersion of fixations along the x-axis in each window
	3.27. mean_fixation_dispersionY: average dispersion of fixations along the y-axis in each window
	3.28. min_fixation_dispersionY: minimum dispersion of fixations along the y-axis in each window
	3.29. max_fixation_dispersionY: maximum dispersion of fixations along the y-axis in each window
	3.30. var_fixation_dispersionY: variation of dispersion of fixations along the y-axis in each window
	3.31. saccade_frequency: frequency of saccades in each window
	3.32. gl_ratio: the ratio between long and short saccades in the window
	3.33. skweness_of_amplitude: skweness of the amplitude of the saccades
	3.34. mean_saccade_duration: average duration of the saccades in the window
	3.35. min_saccade_duration: minimum duration of the saccades in the window
	3.36. max_saccade_duration: maximum duration of saccades in the window
	3.37. var_saccade_duration: variation in the duration of the saccades in the window
	3.38. mean_saccade_length: average length of the saccades in the window
	3.40. min_saccade_length: minimum length of the saccades in the window
	3.41. max_saccade_length: maximum length of the saccades in the window
	3.42. var_saccade_length: variation in length of saccades in window
4. only a subset of these features were used to train the model. This subset includes:
	4.1. condition
	4.2. num_ways
	4.3. delta
	4.4. delta_t
	4.5. all fixation-related features
	4.6. all saccade-related features
5. the model classifies the turning activity and returns the probability of each of the three turning activities naming TL, TR and NT for each of the windows
	