推荐|drop解决过拟合的情况

drop解决过拟合的情况

用到的训练数据集：sklearn数据集

可视化工具：tensorboard，这儿记录了loss值（预测值与真实值的差值），通过loss值可以判断训练的结果与真实数据是否吻合

过拟合：训练过程中为了追求完美而导致问题

过拟合的情况：蓝线为实际情况，在误差为10的区间，他能够表示每条数据。

　　　　　　　橙线为训练情况，为了追求0误差，他将每条数据都关联起来，但是如果新增一些点（+），他就不能去表示新增的点了

训练得到的值和实际测试得到的值相比，训练得到的loss更小，但它与实际不合，并不是loss值越小就越好

drop处理过拟合后:

代码：

import tensorflow as tf
from sklearn.datasets import load_digits
from sklearn.cross_validation import train_test_split
from sklearn.preprocessing import LabelBinarizer

# load data
digits = load_digits()
X = digits.data
y = digits.target
y = LabelBinarizer().fit_transform(y)   # 转换格式
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.3)


def add_layer(inputs, in_size, out_size, layer_name, active_function=None):
    """
    :param inputs:
    :param in_size: 行
    :param out_size: 列 ， [行， 列] =矩阵
    :param active_function:
    :return:
    """
    with tf.name_scope('layer'):
        with tf.name_scope('weights'):
            W = tf.Variable(tf.random_normal([in_size, out_size]), name='W')  #
        with tf.name_scope('bias'):
            b = tf.Variable(tf.zeros([1, out_size]) + 0.1)  # b是一行数据，对应out_size列个数据
        with tf.name_scope('Wx_plus_b'):
            Wx_plus_b = tf.matmul(inputs, W) + b
        Wx_plus_b = tf.nn.dropout(Wx_plus_b, keep_prob=keep_prob)
        if active_function is None:
            outputs = Wx_plus_b
        else:
            outputs = active_function(Wx_plus_b)
        tf.summary.histogram(layer_name + '/outputs', outputs)  # 1.2.记录outputs值，数据直方图
        return outputs


# define placeholder for inputs to network
keep_prob = tf.placeholder(tf.float32)  # 不被dropout的数量
xs = tf.placeholder(tf.float32, [None, 64])  # 8*8
ys = tf.placeholder(tf.float32, [None, 10])

# add output layer
l1 = add_layer(xs, 64, 50, 'l1', active_function=tf.nn.tanh)
prediction = add_layer(l1, 50, 10, 'l2', active_function=tf.nn.softmax)

# the loss between prediction and really
cross_entropy = tf.reduce_mean(-tf.reduce_sum(ys * tf.log(prediction), reduction_indices=[1]))
tf.summary.scalar('loss', cross_entropy)  # 字符串类型的标量张量，包含一个Summaryprotobuf  1.1记录标量(展示到直方图中 1.2 )
# training
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)

sess = tf.Session()
merged = tf.summary.merge_all()  # 2.把所有summary节点整合在一起,只需run一次,这儿只有cross_entropy
sess.run(tf.initialize_all_variables())

train_writer = tf.summary.FileWriter('log/train', sess.graph)  # 3.写入
test_writer = tf.summary.FileWriter('log/test', sess.graph)  # cmd cd到log目录下，启动 tensorboard --logdir=log\

# start training
for i in range(500):
    sess.run(train_step, feed_dict={xs: X_train, ys: y_train, keep_prob: 0.5})  # keep_prob训练时保留50%， 当这儿为1时，代表不drop任何数据，（没处理过拟合问题）
    if i % 50 == 0:
        # 4. record loss
        train_result = sess.run(merged, feed_dict={xs: X_train, ys: y_train, keep_prob: 1})  # tensorboard记录保留100%的数据
        test_result = sess.run(merged, feed_dict={xs: X_test, ys: y_test, keep_prob: 1})
        train_writer.add_summary(train_result, i)
        test_writer.add_summary(test_result, i)

print("Record Finished !!!")

posted on 2018-06-22 14:43 .Tang 阅读( ...) 评论( ...) 编辑收藏

转载于:https://www.cnblogs.com/tangpg/p/9213375.html

文章知识点与官方知识档案匹配，可进一步学习相关知识

Python入门技能树人工智能机器学习工具包Scikit-learn416686 人正在系统学习中

评论记录：