@ [TOC] "zero-based entry of data mining - used car transaction price forecast" baseline implementation
"Zero-based entry of data mining - used car transaction price forecast" Baseline implementation
One after another in front of machine learning, learning about 2 months time, watermelons not read the book, Andrew Ng video did not read. During the Spring Festival began to participate in Alibaba intelligent operation and maintenance algorithm contest , co-operation with partners and small, barely into the semi-finals, feel incompetent, the foundation is not strong enough. A few days ago seen Tianchi held a "zero-based entry of data mining - used car transaction price forecast," the game, and the official Baseline.
Little early to read it again Baseline, feeling very well written, data analysis, EDA, engineering features, parameter adjustment and so on all have, and the official course, the feeling is a treasured book zero-based entry.
I will implement Baseline process of writing down below for your reference.
Fork Baseline
Baseline link here to
the latest version V11.
- Click on the top right corner of the "Fork" button, fork to Tianchi own laboratory space;
- You can modify the title, you can download Notebook to run locally.
Modify data sets
I tried using a common data set Tianchi, unstable, there will be problems can not be read.
Upload the data to your own personal space.
- In the personal space to build subdirectory "zero";
- Upload files to the directory "zero"
- Notebook in the examination file
Modify the submission path
The submission path is also set to just the new directory.
Run Baseline Notebook
The Baseline is very well written, done after the above configuration, it can normally run.
- Run Notebook;
you can click the center button single step, you can click on the right button to execute the entire program.
- Download submission
- To the contest page and submit.
Baseline without modification, probably the result of 680 points.