The Project Folder

Each project expands inside the specific folder specified when it was created. In this section the aarangement of this folder as well as all the generated files will be explained. It will be assumed that the project name is myproject_1 and it is hosted inside the default folder MWMGEN (thus the full path should be C:\Users\[user_name]\Documents\MWMGEN\myproject_1).

classification: folder for storing the generated classifiers.

labels: folder for storing the generated labels.

Mclassification: folder for storing the final classification, product of the majority voting.

results: folder for storing the results of the analysis.

segmentation: folder for storing the segmentations.

settings: folder for storing the project settings.

[project_name].cfg: the CFG-file is used to load this project.

### File: project_name ###

The CONFIG-file (.cfg) is used when the user wish to load a previous project. It is asked when the button Load Project is pressed from the main menu.

### Folder: settings ###

The settings folder is populated when the user presses the button Finalize & Proceed of the New Project window. It contains 4 MAT-files:

• animal_groups.csv: file containing the animal groups specified by the user. This file can be load when assigning animal groups.

• animal_groups.mat: holds the animal groups specified by the user.

• my_trajectories.mat: holds information about the loaded trajectories.

• new_properties.mat: holds the experiment settings and properties specified by the user.

• tags.txt: holds information about the specified strategies of this project (abbreviation, name, id for distinguish each strategy and color, linestyle for plotting it).

### Folder: segmentation ###

The segmentation folder is populated when the user presses the button Segmentation of the main menu. It contains only MAT-files (.mat) and their number depends on the number of different segmentations. Each file has a different name specified by the segmentation settings: segmentation _ configs _ [number of segments] _ [segments length] _ [segments overlap].mat.

### Folder: labels ###

The segmentation folder is populated when the user presses the button Save of the Browse Trajectories window. It contains equal amount of MAT-files (.mat) and CSV-files (.csv) and their number depends on the number of different labelling that was performed. Each duo of MAT-file and CSV-file has a different name specified by the segmentation settings and the number of labels given for that segmentation: labels _ [number of labels] _ [segments length] _ [segments overlap] [optional: -user note].mat.

### Folder: classification ###

The classification folder contains subfolder which host a number of MAT-files (.mat) depending on the classification setting used to generate the classifiers. It is populated by either pressing the button Default of the main menu or the button Generate Classifiers of the Classification (Advanced) window. Each subfolder has a different name specified by the number of labels given, and the segmentation settings: class _ [number of labels] _ [number of segments] _ [segments length] _ [segments overlap] [optional: -user note]. Each classifier, MAT-file (.mat), has again a distinctive name: classification _ configs _ [number of segments] _ [number of labels] _ [default number of clusters].mat

### Folder: Mclassification ###

The classification folder contains subfolder which host an equal number of MAT-files (.mat) and TEXT-files (.txt) depending on the classification setting used to generate the final classification (merged results of the different classifiers). It is populated by either pressing the button Default of the main menu or the button Merge of the Classification (Advanced) window. Each subfolder has a distinctive name: class _ [number of labels] _ [number of segments] _ [segments length] _ [segments overlap] _ [number of classifiers used] _ [number of iterations] _ [merging rule] [optional: -user note].mat.

merging rule

• Majority voting: mr [value of threshold].

Each MAT-files (.mat) inside each subfolder is named after: merged _ [number].mat. Each equivalent TEXT-files (.txt) is named after: unlabelled _ segs _ [number].txt. In both cases [number] starts from 1 and ends at the [number of iterations].

The format of the TEXT-files (.txt) depends on the [merging rule] used:

merging rule = Majority voting:

• Line 1: [number of segments] _ [list of classifiers used, identified by their default number of clusters, separated with dot '.'].

• Line 2: total segments marked as undefined: 'Total: [total number]'.

• Line 3: total segments marked as undefined because of equal votes between two or more strategies: Undecided: [undecided number].

• Line 4+: A list of all the segments of [Line 3].

• Line 602: In case of threshold more than zero the number of clusters marked as undefined because of the threshold: Threshold: [a number].

• Line 603+: A list of all the segments of [Line 602]

### Folder: results ###