《python数据挖掘入门与实践》第五章的错误

[了然]

代码：

import os
import pandas as pd
data_folder = os.path.join(os.path.expanduser("~"),"data","Ads")
data_filename = os.path.join(data_folder,"ad.data")
def convert_number(x):
    try:
        return float(x)
    except ValueError:
        return np.nan
from collections import defaultdict
converters = defaultdict(convert_number)
converters[1558] = lambda x:1 if x.strip() == "ad." else 0
ads = pd.read_csv(data_filename,header=None,converters=converters)
ads[:5]
x = ads.drop(1558,axis=1).values
y = ads[1558]
from sklearn.decomposition import PCA
pca = PCA(n_components=5)
xd = pca.fit_transform(x)
import numpy as np
np.set_printoptions(precision=3,suppress=True)
pca.explained_variance_ratio_

错误信息：

ValueError                                Traceback (most recent call last)
<ipython-input-10-f726f2ff6f29> in <module>()
      1 from sklearn.decomposition import PCA
      2 pca = PCA(n_components=5)
----> 3 xd = pca.fit_transform(x)
      4 import numpy as np
      5 np.set_printoptions(precision=3,suppress=True)

/home/kongnian/anaconda3/lib/python3.5/site-packages/sklearn/decomposition/pca.py in fit_transform(self, X, y)
    239 
    240         """
--> 241         U, S, V = self._fit(X)
    242         U = U[:, :self.n_components_]
    243 

/home/kongnian/anaconda3/lib/python3.5/site-packages/sklearn/decomposition/pca.py in _fit(self, X)
    266             requested.
    267         """
--> 268         X = check_array(X)
    269         n_samples, n_features = X.shape
    270         X = as_float_array(X, copy=self.copy)

/home/kongnian/anaconda3/lib/python3.5/site-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, warn_on_dtype, estimator)
    371                                       force_all_finite)
    372     else:
--> 373         array = np.array(array, dtype=dtype, order=order, copy=copy)
    374 
    375         if ensure_2d:

ValueError: could not convert string to float: '?'

数据集：
http://archive.ics.uci.edu/ml/datasets/Internet+Advertisements download Advertisements dataset
操作系统信息：
Linux ubuntu 4.4.0-40-generic #60-Ubuntu SMP Fri Sep 23 16:45:45 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

[斯温jack]

可能是你的数据预处理问题 预处理后的数据有字符串。

[了然]

此问题已解决

[baidu_31803731]

楼主怎么解决的，求解

[zzz47896]

怎么解决的啊

[枭亭]

我也遇到了同样的问题，求解决方法

[sammy_cc]

引用 1 楼 sinat_30665603 的回复:

可能是你的数据预处理问题 预处理后的数据有字符串。

请教一下如何解决的

[姜红杰]

人呢,咋解决的

[flybird101]

我的做法： from collections import defaultdict converters = defaultdict(convert_number) for i in ads.columns[:-1]: converters[i] = convert_number converters[1558] = lambda x: 1 if x.strip() == "ad." else 0 ads = pd.read_csv(data_filename, header=None, converters=converters) ads = ads.dropna(axis=0, how='any') ads[:100]

[qq_40837600]

原因：数据集内有含有空格的字符 解决方案： #将Dataframe中的所有异常字符换成nan（参照报错信息） ads = ads.replace("?",np.nan).replace(" ?",np.nan).replace(" ?",np.nan) 再次转化数据类型 ads = ads.astype(float).dropna() 确认数据： ads.dtypes #转换成功