{"cells":[{"metadata":{},"cell_type":"markdown","source":"# The example of calculating Global Average Precision using the Evaluation library   \nhttps://github.com/yisaienkov/evaluations   \nhttps://evaluations.readthedocs.io/en/latest/","execution_count":null},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# Internet ON.\n\n# !pip install -U pip\n# !pip install evaluations\n\n\n# Internet OFF. \n# You can use evaluations dataset (see input folders) (the same as in pypi).\n\n!pip install --no-deps '../input/evaluations/'","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"import random\n\nfrom evaluations.kaggle_2020 import global_average_precision_score\nimport pandas as pd","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"df_train = pd.read_csv('../input/landmark-recognition-2020/train.csv', index_col=0)\ndf_train","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### True labels","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"correct_labels = df_train.to_dict()['landmark_id']","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Fully corrected predictions","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"predicted_labels = {key: (item, random.random()) for key, item in correct_labels.items()}","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Score with corrected predictions = 1","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"global_average_precision_score(correct_labels, predicted_labels)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Every second prediction is correct","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"predicted_labels = {}\nfor ind, (key, item) in enumerate(correct_labels.items()):\n    if ind % 2:\n        predicted_labels[key] = (item, random.random())\n    else:\n        predicted_labels[key] = (1, random.random())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Score is not good","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"global_average_precision_score(correct_labels, predicted_labels)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Fully random predictions","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"labels = df_train['landmark_id'].unique().tolist()\npredicted_labels = {key: (random.sample(labels, 1)[0], random.random()) for key, item in correct_labels.items()}","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Bad score","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"global_average_precision_score(correct_labels, predicted_labels)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}