Skip to content

Commit e6e9970

Browse files
authored
Update Report.md
1 parent 353d6da commit e6e9970

File tree

1 file changed

+7
-9
lines changed

1 file changed

+7
-9
lines changed

LinearClassifiers/Report.md

Lines changed: 7 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@
44

55
2、 分别使用感知器算法学习区分下列两类样本的线性分类器:
66

7-
![img](.\sample_data.png)
7+
![img](./sample_data.png)
88

99
3、 MNIST数据集测试:使用TrainSamples中的30000个17维特征手写数字样本训练线性分类器区分10个类别,TrainLabels中包含训练样本的标签;测试线性分类器对TestSamples中10000个样本的识别正确率。
1010

@@ -31,10 +31,9 @@ F(x1, x2) = 0.4469281*x1 + 0.185913*x2 – 0.6
3131
![img](./multiclass.png)
3232

3333
- 分类正确率:
34-
- 感知器算法:
35-
36-
```
37-
precision recall f1-score support
34+
- 感知器算法:
35+
36+
precision recall f1-score support
3837
0 0.902 0.888 0.895 1025
3938
1 0.917 0.894 0.905 1179
4039
2 0.767 0.838 0.801 882
@@ -45,10 +44,9 @@ F(x1, x2) = 0.4469281*x1 + 0.185913*x2 – 0.6
4544
7 0.876 0.864 0.870 1061
4645
8 0.758 0.726 0.741 1007
4746
9 0.734 0.777 0.755 940
48-
49-
avg / total 0.821 0.820 0.820 10000
50-
Test acc: 0.8196
51-
```
47+
48+
avg / total 0.821 0.820 0.820 10000
49+
Test acc: 0.8196
5250

5351

5452
- 最小平方误差算法:

0 commit comments

Comments
 (0)