Question

我有一套540的训练集和150个图像像素的测试集。这些值存储在不同的csv文件中，如下所示：

[label],[num0],[num1],...,[num399]

标签是单个字母，400个是像素值。该套装用于手语识别。

代码 -

import numpy as np 
import os
import csv
from sklearn import svm
from sklearn import cross_validation
from sklearn import linear_model

path = '/home/goel/skin'


X_train=[]
y_train=[]
X_test=[]
y_test=[]
ylist=[]

with open("20_20_centered_newer.csv",'r') as file:
    reader = csv.reader(file,delimiter=',')
    reader.next()
    for row in file:
        y_train.append(row[0])
        if row[0] not in ylist:
            ylist.append(row[0])        
        row=row[2:]
        row=[int(x) for x in row.split(',')]
        X_train.append(np.array(row))

y2list=[]

with open("20x20_test.csv",'r') as file:
    reader = csv.reader(file,delimiter=',')
    for row in file:
        y_test.append(row[0])
        if row[0] not in y2list:
            y2list.append(row[0])       
        row=row[2:]
        row=[int(x) for x in row.split(',')]
        X_test.append(np.array(row))

print ylist
print y2list

#clf = linear_model.SGDClassifier().fit(X_train,y_train)
#clf = svm.SVC(kernel='linear').fit(X_train,y_train)
#clf = svm.LinearSVC().fit(X_train,y_train)
clf = linear_model.LogisticRegression().fit(X_train,y_train)
print clf.score(X_test,y_test)

显然，我在所有分类器中得到的分数为.78，最多12位小数！

这里是否存在我不知道的语义错误？

在不同的分类器中获得相同的准确度 - sklearn

0 个答案: