Dataset For Application Of Recommender Systems On Police Photo Lineup Assembling Task

For more information about the Recommender Systems for Police Photo Lineup project, please visit: http://www.ksi.mff.cuni.cz/~peska/lineup/paper.pdf The dataset consists of two parts: <strong>The raw dataset </strong>contains visual and attribute-based desriptors of the list of candidate persons: - personsCB_IDs.csv: ordered list of persons IDs<br> - personsData.csv: raw attribute features of the persons<br> - personsCBVectors.csv: derived binary attribute-features with TF-IDF applied. The ordering of records is the same as in personsCB_IDs.csv<br> - personsVectors.csv: derived visual descriptors of persons' images. Probability layer of VGG-Face CNN was applied for this task. The ordering of records is the same as in personsCB_IDs.csv. <strong>The implicit feedback dataset</strong> (feedbackDatasetCB-RSFeatures.csv) contains information received from the user-study on lineup assembling task performed by seven domain experts. This table has following structure: - evaluatorID;<br> - lineupID (id of the suspect);<br> - candidateID (person recommended to this particular suspect);<br> - calculated content-based similarity<br> - selection (1 if evaluator selected this candidate, 0 otherwise);<br> - substraction of CB features of the suspect and the candidate (this can be simply modified by accessing raw dataset to contain visual descriptors or any other combination of features)