模型的特征数量必须与输入匹配吗? [英] Number of features of the model must match the input?
问题描述
我正在尝试对我拥有的一些数据使用 RandomForestClassifier.代码如下:
I'm trying to use a RandomForestClassifier on some data I have. The code is below:
print train_data[0,0:20]
print train_data[0,21::]
print test_data[0]
print 'Training...'
forest = RandomForestClassifier(n_estimators=100)
forest = forest.fit( train_data[0::,0::20], train_data[0::,21::] )
print 'Predicting...'
output = forest.predict(test_data)
但这会产生以下错误:
ValueError:模型的特征数必须与输入匹配.模型 n_features 为 3,输入 n_features 为 21
ValueError: Number of features of the model must match the input. Model n_features is 3 and input n_features is 21
前三个打印语句的输出是:
The output from the first three print statements is:
[ 0. 0. 0. 0. 1. 0.
0. 0. 0. 0. 1. 0.
0. 0. 0. 37.7745986 -122.42589168
0. 0. 0. ]
[ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
1. 0.]
[ 0. 0. 0. 0. 0. 0.
0. 1. 0. 0. 1. 0.
0. 0. 0. 0. 37.73505101
-122.3995877 0. 0. 0. ]
我假设数据对于我的 fit
/predict
调用是正确的格式,但它在 predict
上出错.谁能看到我在这里做错了什么?
I had assumed that the data was in the correct format for my fit
/predict
calls, but it is erroring out on the predict
. Can anyone see what I am doing wrong here?
推荐答案
用于训练模型的输入数据是train_data[0::,0::20]
,我认为这是一个错误(为什么跳过中间的功能?)——它应该是 train_data[0::,0:20]
而不是基于你在开始时所做的调试打印.
The input data used to train the model is train_data[0::,0::20]
, which I think is a mistake (why skip features in between?) -- it should be train_data[0::,0:20]
instead based on the debug prints you did in the beginning.
此外,似乎最后一列代表了 train_data
和 test_data
中的标签.在预测时,您可能希望在调用 predict
函数时传递 test_data[:, :20]
而不是 test_data
.
Also, it seems that the last column represents the labels in both train_data
and test_data
. When predicting, you might want to pass test_data[:, :20]
instead of test_data
when calling thepredict
function.
这篇关于模型的特征数量必须与输入匹配吗?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!