After learning the basic knowledge of deep learning, participating in practice is the best way to continue to improve yourself.

On the "Overview" tab of each competition, you can see some information about the competition and its data set, the evaluation criteria for submitting valid results (each competition is slightly different), and the frequently asked questions of the competition. On the Data tab, you can see a short description of the data.

What you need are these three files: train.csv, test.csv and data_description.txt Please put them in a folder that you can access quickly. The "Discussion" tab is like an exclusive forum for competitions. In popular competitions, these discussions often contain very valuable information, because the terms of the competition sometimes require participants to disclose any information they use on the discussion board.

Data leakage is difficult to avoid and deal with, and it happens occasionally in competitions. On the one hand, making full use of data can get higher scores to win the competition; On the other hand, the model combined with data leakage is usually useless to practice, so it is not supported by the competition. Diligent participants often share data leaks on discussion boards to help the competitive environment become better.

relevant information

Kaggle was founded by co-founder and CEO Anthony Goldbloom in Melbourne in 20 10, which mainly provides a platform for developers and data scientists to hold machine learning competitions, host databases, and write and share codes. The platform has attracted the attention of 800,000 data scientists, and these user resources may be the main factor to attract Google.

Kaggle was founded in Melbourne in 20 10 by co-founder and CEO Anthony Goldbloom. It mainly provides a platform for developers and data scientists to hold machine learning competitions, host databases, write and share codes. This platform has attracted the attention of many scientists and developers, and they have also settled in this platform.