关于少量数据的svm、svr和logistics的分类比较

最新推荐文章于 2023-11-20 10:19:40 发布

yinghua2016

最新推荐文章于 2023-11-20 10:19:40 发布

阅读量812

点赞数

本文链接：https://blog.csdn.net/yinghua2016/article/details/51954340

版权

本文通过代码展示了在少量数据情况下，使用SVM、SVR和逻辑回归进行分类的实现过程。通过对数据集'cloth_feature.txt'和'model_feature.txt'的处理，构建输入特征并进行训练，最终用classification_report评估了模型的性能。

摘要由CSDN通过智能技术生成

SVM

from sklearn.svm import SVC
from sklearn.metrics import classification_report

def read_file(file_name):
with open(file_name) as fp:
vectors = []
for line in fp:
items = line.split('\t')
values = []
for item in items:
try:
values.append(float(item))
except:
continue
vectors.append(values)

return vectors

def connect_vector(file1, file2):
cloth_vectors = read_file(file1)
model_vectors = read_file(file2)

X = []
vector = []
for v1 in cloth_vectors:
for v2 in model_vectors:
vector.append(v1)
vector.append(v2)
X.append(v1+v2) #4096*2

X1 = []
X11= []
X2 = []
X22= []
y1 = []
y11= []

for i in range(0,1024,32):
X1.append(X[i]);
for i in range(0,16,1):
X11.append(X1[i])
for i in range(1,17,1):
X11.append(X[i])
X1.append(X[33])

for i in range(0,1024,32):
X2.append(X[i]);
for i in range(16,32,1):
X22.append(X2[i])
for i in range(991,1007,1):
X22.append(X[i])