项目作者: sethuiyer

项目描述 :
This code attempts to visualise the Game of thrones dataset in kaggle using t-SNE Visualization
高级语言: Python
项目地址: git://github.com/sethuiyer/visualize-GOT.git
创建时间: 2016-12-24T08:34:50Z
项目社区:https://github.com/sethuiyer/visualize-GOT

开源协议:

下载


visualize-GOT

This code attempts to visualise the Game of thrones dataset in kaggle, Particularly T-SNE Visualization

Question 1:

To which class is the NaN of the battles[last_row][attacker_outcome_column] is similar to? Win or Lose?

Executing python visualise_battle.py, we have this following graph

attacker

So, Mostly it is similar to the win outcome.

Question 2:

How allegiances, nobility and the appearence in the book affect the gender of charecter deaths.
figure

We see some outliers in the death of male charecters.

Question 3:

Which feature is more powerful in the prediction? Popularity or the fact that charecter is actually alive or not?

figure

Here, we see clusters are forming among the classes which have the common theme of popularity. Hence, we can say popularity is much more a powerful feature in the prediction.