Discussion of article "Deep Neural Networks (Part III). Sample selection and dimensionality reduction"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
New article Deep Neural Networks (Part III). Sample selection and dimensionality reduction has been published:
This article is a continuation of the series of articles about deep neural networks. Here we will consider selecting samples (removing noise), reducing the dimensionality of input data and dividing the data set into the train/val/test sets during data preparation for training the neural network.
Let us conduct an experiment. Let us treat the following data sets: DT (raw data before preliminary processing), DTn (only normalized raw data set), DTTanh.n (without outliers, tan-transformed and normalized) and the train set with the ORBoostFilter() function, which is a filter for removing noise. Let us see how the distribution changed after such processing.
Fig. 1. Distribution of predictors in the sets after removing noise samples
Author: Vladimir Perervenko