本实验分三步:

1. 建立数据集

2. 建立网络并训练

3. 可视化

import numpy as np
from keras.models import Sequential
from keras.layers import Dense
from keras.optimizers import SGD # 构建数据集
X_data = np.linspace(-1,1,100)[:, np.newaxis]
noise = np.random.normal(0,0.05,X_data.shape)
y_data = np.square(X_data) + noise + 0.5 # 构建神经网络
model = Sequential()
model.add(Dense(10, input_dim=1, init='normal', activation='relu'))
model.add(Dense(1, init='normal'))
sgd = SGD(lr=0.1)
model.compile(loss='mean_squared_error', optimizer=sgd)
# 训练
model.fit(X_data, y_data, nb_epoch=1000, batch_size=100, verbose=0)
# 在原数据上预测
y_predict=model.predict(X_data,batch_size=100,verbose=1) # 可视化
import matplotlib.pyplot as plt
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.scatter(X_data, y_data)
ax.plot(X_data,y_predict,'r-',lw=5)
plt.show()

训练结果:

aaarticlea/png;base64,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" alt="" width="523" height="356" />

keras 入门之 regression的更多相关文章

  1. Keras入门(四)之利用CNN模型轻松破解网站验证码

    项目简介   在之前的文章keras入门(三)搭建CNN模型破解网站验证码中,笔者介绍介绍了如何用Keras来搭建CNN模型来破解网站的验证码,其中验证码含有字母和数字.   让我们一起回顾一下那篇文 ...

  2. keras 入门整理 如何shuffle,如何使用fit_generator 整理合集

    keras入门参考网址: 中文文档教你快速建立model keras不同的模块-基本结构的简介-类似xmind整理 Keras的基本使用(1)--创建,编译,训练模型 Keras学习笔记(完结) ke ...

  3. Keras入门(二)模型的保存、读取及加载

    本文将会介绍如何利用Keras来实现模型的保存.读取以及加载.   本文使用的模型为解决IRIS数据集的多分类问题而设计的深度神经网络(DNN)模型,模型的结构示意图如下: 具体的模型参数可以参考文章 ...

  4. Keras入门(一)搭建深度神经网络(DNN)解决多分类问题

    Keras介绍   Keras是一个开源的高层神经网络API,由纯Python编写而成,其后端可以基于Tensorflow.Theano.MXNet以及CNTK.Keras 为支持快速实验而生,能够把 ...

  5. Keras入门(六)模型训练实时可视化

      在北京做某个项目的时候,客户要求能够对数据进行训练.预测,同时能导出模型,还有在页面上显示训练的进度.前面的几个要求都不难实现,但在页面上显示训练进度当时笔者并没有实现.   本文将会分享如何在K ...

  6. 数据挖掘入门系列教程(十一)之keras入门使用以及构建DNN网络识别MNIST

    简介 在上一篇博客:数据挖掘入门系列教程(十点五)之DNN介绍及公式推导中,详细的介绍了DNN,并对其进行了公式推导.本来这篇博客是准备直接介绍CNN的,但是想了一下,觉得还是使用keras构建一个D ...

  7. 深度学习:Keras入门(一)之基础篇

    1.关于Keras 1)简介 Keras是由纯python编写的基于theano/tensorflow的深度学习框架. Keras是一个高层神经网络API,支持快速实验,能够把你的idea迅速转换为结 ...

  8. 深度学习:Keras入门(二)之卷积神经网络(CNN)

    说明:这篇文章需要有一些相关的基础知识,否则看起来可能比较吃力. 1.卷积与神经元 1.1 什么是卷积? 简单来说,卷积(或内积)就是一种先把对应位置相乘然后再把结果相加的运算.(具体含义或者数学公式 ...

  9. keras入门(三)搭建CNN模型破解网站验证码

    项目介绍   在文章CNN大战验证码中,我们利用TensorFlow搭建了简单的CNN模型来破解某个网站的验证码.验证码如下: 在本文中,我们将会用Keras来搭建一个稍微复杂的CNN模型来破解以上的 ...

随机推荐

  1. Python - twisted web 入门学习之一

    原文地址:http://zhouzhk.iteye.com/blog/765884 python的twisted框架中带了一个web server: twisted web.现在看看怎么用. 一)准备 ...

  2. C# Request中修改header信息

    var headers = app.Context.Request.Headers; Type hdr = headers.GetType(); PropertyInfo ro = hdr.GetPr ...

  3. C# 中 多线程同步退出方案 CancellationTokenSource

    C# 中提供多线程同步退出机制,详参对象: CancellationTokenSource CancellationTokenSource 中暂未提供复位操作,因此当调用Cancle 之后,若再次调用 ...

  4. 获取数据库表详细信息、存储过程、视图、的sql

    select s.[name] + '.' + t.[name] as tablename from sys.tables as t,sys.schemas as s where t.schema_i ...

  5. Common.Logging log4net Common.Logging.Log4Net 配置

    1.log4net 单独配置 log4net支持多种格式的日志输出,我这里只配置输出到本地的txt文件这种格式. <log4net> <root> <appender-r ...

  6. hbase数据迁移-HDFS拷贝

    1.把数据表test从hbase下拷出 hdfs dfs -get /hbase/data/default/test /home/hadoop/hbase/test 2.文件放到新集群的系统上 scp ...

  7. 重置了下系统好多关于mysql密码的文章都很渣拷分好的备用

    方法1: 用SET PASSWORD命令 mysql -u root mysql> SET PASSWORD FOR 'root'@'localhost' = PASSWORD('newpass ...

  8. HighChats图表控件显示精度小数点的方法

    相信大家对highchats这个图表控件并不陌生,最近在项目中用到它,但是某些字段需要显示为小数点,顾查找资料文档发现下面两个方式可以实现: 初始化时候添加如下两个参数 tooltip:{    fo ...

  9. C# FTP/SFTP文件传输控件FTP for .NET/.NET CF 详细介绍

    FTP for .NET将FTP客户端功能添加到您的应用程序之中..NET控件的FTP支持所有常用的FTP服务器以及代理服务器,包括可扩展的目录解析.同步以及异步操作.主动与被动模式.以VB.NET与 ...

  10. nim也玩一行流,nim版的list comprehension

    nim 是一门风格类似python的静态编译型语言,官方网站:http://nim-lang.org 如果你想折腾nim的编辑环境,可以用sublime text3 +插件nimlime,notepa ...