Dynamic Deep Learning Algorithm (DDLA) for Processing of Complex and Large Datasets

Abstract

Nowadays, deep learning (DL) is blueprint for the machine learning (ML). Compare with NL, DL is most efficient, time and with low cost There are no limitations for the learning approaches in DL. DL algorithms can extract high-quality of features based on the dataset provided. Due to the fast development in internet technology, huge data is started processing by dividing according to the requirements. DL algorithms are more compatible for processing of huge and complex datasets such as pattern matching, recognition of handwriting, recognition of speeches, analysis of stock markets and many more. DL has more advancement on various applications and this will solve various issues in complicated pattern applications. Previously, datasets such as handwritten is used to find the accurate result with the Ensemble novel classifier (ENC). But this not worked up to the mark. In this paper, dynamic deep learning algorithm (DDLA) is designed and developed to process the complicated and complex datasets. This is the combination of Auto Encoder (AE) and Adaptive Convolutional neural network (A-CNN). Experimental results show the performance of the EDDLA with ENC.