BruunoMM / spotify-skip-prediction-top-1-solution

Our Codes for the WSDM Cup Skip Prediction Challenge (1st Position on the Final Leaderboard)

This repository contains the Python source code of our solutions to the WSDM Cup 2019.

Instructions for Running the Code

Data Preparation

• Firstly, create three folders train_fold, test_fold, and Skip_Data, and put the decompressed train csv files, test csv files, and song feature files there;
• Create a new folder test_pred for storing the processed test files
• Run the Spotify_Skip_Final_Solution_Data_Preparation script, note that this script consists of 4 segments, before running the 4th segment, please use the official Glove code at https://github.com/stanfordnlp/GloVe for learning the embeddings from the glove_data.txt file generated in previous steps. Glove training parameters areVOCAB_MIN_COUNT=7, VECTOR_SIZE=150, MAX_ITER=150 ,WINDOW_SIZE=10. This part is quite time consuming, so you may also access the trained embedding from: https://pan.baidu.com/s/1bhJbsGSD7bXf4txmsqaplA

Model Fitting

Create a new folder Data for storing trained model. Run the Spotify_Skip_Final_Solution_Model_Fitting script.

Submission Generation

Run the Spotify_Skip_Final_Solution_Test_File_Prediction and then the Spotify_Skip_Final_Solution_Submission_Generation scripts. The generated submission file should score 0.648 and achieve top 1 on the leaderboard. To match the 0.651 result, you may only need to train the model 5-6 times, and then average/ensemble the prediction results (Our actual ensembling approach used 6 slight modifications of the core model defined in Spotify_Skip_Final_Solution_Model_Fitting script, for example by replacing ReLU activations with ELU activations. But such modifications may not be necessary. Please contact me if you need exact definitions of these model variants).

License

Copyright [2019][Lin Zhu]

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Contact

Our scripts are a little messy and we think there must be some potential mistakes in our codes. Please feel free to contact us, if there are any question about our project or suggestion to help improve. email: [e-mail]