turicreate学习笔记2-clustering turicreate学习笔记2-clustering

1.生成sframe文件

import turicreate as tcsf_functional = tc.SFrame.read_csv('train/train_FNC.csv') sf_morphometry = tc.SFrame.read_csv('train/train_SBM.csv')sf = sf_functional.join(sf_morphometry, on='Id') sf = sf.remove_column('Id')sf.save('schizophrenia_clean.sframe')

【turicreate学习笔记2-clustering】train_FNC.csv和train_SBM.csv在kaggle上可以下载

文章图片
sframe file 2.create kmeans model

# create kmeans model from math import sqrtK = int(sqrt(sf.num_rows() / 2.0))kmeans_model = tc.kmeans.create(sf, num_clusters=K) print kmeans_model.summary()

文章图片
kmeans model 3.使用

kmeans_model['cluster_info'].print_rows( num_columns=5, max_row_width=80, max_column_width=10 )new_clusters = kmeans_model.predict(sf[:5]) print new_clusters