[英]project PCA back into original scales with explained_variance_ratio_ condition
I have 2 questions concerning PCA when using scikit. 使用scikit时,我有2个关于PCA的问题。
Lets suppose I have the following data: 假设我有以下数据:
fullmatrix =[[2.5, 2.4],
[0.5, 0.7],
[2.2, 2.9],
[1.9, 2.2],
[3.1, 3.0],
[2.3, 2.7],
[2.0, 1.6],
[1.0, 1.1],
[1.5, 1.6],
[1.1, 0.9]]
Now I do the PCA calculations: 现在,我进行PCA计算:
from sklearn.decomposition import PCA as PCA
sklearn_pca = PCA()
Y_sklearn = sklearn_pca.fit_transform(fullmatrix)
print Y_sklearn # Y_sklearn is now the Data transformed with 2 eigenvectors
sklearn_pca.explained_variance_ratio_ # variance explained by each eigenvector
print sklearn_pca.explained_variance_ratio_
sklearn_pca.components_ # eigenvectors order by highest eigenvalue
print sklearn_pca.components_
First question: How can I project back this Y_sklearn into the original scale? 第一个问题:如何将这个Y_sklearn投影回原始比例? (I know we should get back the same data as of full matrix as I'm using all eigenvectors, its just to check if done right).
(我知道我们应该使用所有特征向量来获取与全矩阵相同的数据,只是为了检查是否正确)。
Second question: How can I enter a threshold regarding minimum acceptable total variance coming from "sklearn_pca.explained_variance_ratio_"?. 第二个问题:如何输入有关“ sklearn_pca.explained_variance_ratio_”的最小可接受总方差的阈值? For example lets say I want to keep using eigenvectors until when i reach total explained_variance_ratio_ above 95%.
例如,假设我要一直使用特征向量,直到达到95%以上的总explained_variance_ratio_。 In this case is easy, we just use the first eigenvector as it explains .96318131%.
在这种情况下很容易,我们只使用第一个特征向量即可,其解释为0.996318131%。 But how can we do this in a more automated way?
但是,我们如何才能以更自动化的方式做到这一点呢?
First: sklearn_pca.inverse_transform(Y_sklearn)
首先:
sklearn_pca.inverse_transform(Y_sklearn)
Second: 第二:
thr = 0.95
# Is cumulative sum exceeds some threshold
is_exceeds = np.cumsum(sklearn_pca.explained_variance_ratio_) >= thr
# Which minimal index provides such variance
# We need to add 1 to get minimum number of eigenvectors for saving this variance
k = np.min(np.where(is_exceeds))+1
# Or you can just initialize your model with thr parameter
sklearn_pca = PCA(n_components = thr)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.