Auto Encoder Fixed-Target Training Features Extraction Approach for Binary Classification Problems

The main issues with machine learning-based feature extraction techniques are the requirement of extensive domain-level knowledge, experience, and need to be supported by large amounts data that sometimes not available. Moreover, it is often difficult apply knowledge extract necessary features for building a machine-learning classifier. Therefore, significantly important find develop depend mainly on training don’t require or experience. To address these binary classification problems, novel approach, AE-FT(Fixed Target) extracting common using Deep Belief Network (DBN)-based Autoencoder (AE) proposed in this paper. In extracted DBN trained dataset sample’s Fixed Target approach.
 approach tested evaluated two different sets. For each dataset, used train seven learning algorithms compared their performances. number very small other existing methods. method improves performance reducing laborious processes, increasing recognition accuracy effectively.
 results show without any human expertise, provides good techniques.