Familiarity-dependent computational modelling of indoor landmark selection for route communication: a ranking approach

These codes and data are aimed at computing suitable indoor landmarks for users of different familiarity. The descriptions of files are as follows:<b>rel_ranking_familiarity.py</b>: the source code to train models;<b>hyperparams</b>: this fold includes <i>lgbrank_familiar_best_params.csv</i> and <i>lgbrank_unfamiliar_best_params.csv </i>files, which store the hyperparameters for 17 fold cross_validation in the familiar datase and in the unfamiliar dataset, respectively;<b>17_fold_dataset</b>: this fold includes all the dataset for 1-_fold_cross_validation, in each fold <i>familiar_rank.train.query</i> and <i>familiar_rank.train</i> are for training in the familiar dataset,<i> unfamiliar_rank.train.query </i>and<i> unfamiliar_rank.train</i> are for training in the unfamiliar dataset, <i>familiar_rank.test.query </i>and <i>familiar_rank.test </i>are for testing in the familiar dataset,<i> unfamiliar_rank.test.query</i> and <i>unfamiliar_rank.test </i>are for testing in the unfamiliar dataset.<br><br>