机器学习
Nirvana;
图像算法,机器视觉,AI算法方向
展开
-
【Python】Dropout抑制过拟合
Tensorflow实现Dropout抑制过拟合import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# Dropout抑制过拟合(train_image, train_label), (test_image, test_label) = tf.keras.datasets.fashion_mnist.load_data()train_label_onehot =原创 2021-12-30 17:31:18 · 240 阅读 · 1 评论 -
【Python】过拟合问题
Tensorflow实现抑制过拟合import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# 过拟合问题(train_image, train_label), (test_image, test_label) = tf.keras.datasets.fashion_mnist.load_data()train_label_onehot = tf.keras.utils原创 2021-12-30 17:29:34 · 251 阅读 · 1 评论 -
【Python】模型优化与超参数选择
Tensorflow实现模型优化与超参数选择import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# 模型优化与超参数选择(train_image, train_label), (test_image, test_label) = tf.keras.datasets.fashion_mnist.load_data()train_label_onehot = tf.k原创 2021-12-30 17:25:38 · 292 阅读 · 0 评论 -
【Python】逻辑回归(softmax多分类)
Tensorflow实现逻辑回归-softmax多分类import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# softmax多分类(train_image, train_label), (test_image, test_label) = tf.keras.datasets.fashion_mnist.load_data()print(train_image.s原创 2021-12-24 16:04:59 · 1702 阅读 · 0 评论 -
【Python】多层感知机
import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# 多层感知机data = pd.read_csv('Advertising.csv') # 预测广告投放与效果 多对一 多包括电视、广播和报纸print(data.head())plt.scatter(data.TV, data.sales)plt.show()x = data.iloc[:, 1:-原创 2021-12-24 15:53:09 · 1786 阅读 · 0 评论 -
【Python】线性回归
TensorFlow实现线性回归import tensorflow as tfimport pandas as pdimport matplotlib.pyplot as plt# 线性回归print(tf.__version__)# 数据集 - 受教育年限与收入的关系data = pd.read_csv('Income1.csv')x = data.Educationy = data.Incomeprint(data)# 数据可视化plt.scatter(x, y)# plt原创 2021-12-22 18:20:48 · 1101 阅读 · 0 评论 -
【Python】逻辑回归(二分类)
Tensorflow实现逻辑回归-二分类import tensorflow as tfimport pandas as pdimport numpy as npimport matplotlib.pyplot as plt# 逻辑回归 - 二分类data = pd.read_csv('credit-a.csv', header=None)print(data.iloc[:, -1].value_counts())x = data.iloc[:, :-1]y = data.iloc[:原创 2021-12-24 15:58:08 · 1211 阅读 · 1 评论