Ytukyg

I'm using normalized data ranging [-1,1]. I have to used 1 dimensional convolution. But my model is only showing 74 percent training and test accuracy. I want to increase this accuracy. I have increases the amount of input data and and tried to add 2 convolution layer but still there is no effect on accuracy. I have to classify data into two classes. Can someone please guide me what should I do to increase accuracy. Or suggest some CNN network that is using data [-1,1]. Moreover should I add more fully connected layer or should I increase convolutional layers which will be more effective?

input = tf.reshape(x, [-1, FLAGS.image_width, FLAGS.input_channel])

filter = weight_variable([FLAGS.filter_width,  FLAGS.input_channel,

FLAGS.filter_channel])



conv_out = tf.nn.tanh(conv1d(input, filter))

pool_out = pool(conv_out2)

dim = pool_out.get_shape().as_list()



conv_re = tf.reshape(pool_out, (-1, dim[1]*dim[2]))

W_fc = weight_variable([dim[1]*dim[2], 2])

logits = tf.matmul(conv_re, W_fc)

y_prime = tf.nn.softmax(logits)



cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits, 

labels= y_)

loss = tf.reduce_mean(cross_entropy)

optimizer = tf.train.GradientDescentOptimizer(FLAGS.rLearn).minimize(loss)