.png){kind=icon}
.png){kind=icon}
Return
Bunch graph based dimensionality reduction using auto-encoder for character recognition
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-022-12907-y.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Sometimes a group of similar types of dimensions is also treated as nodes. However, these groups can be considered as bunch nodes which may contain several nodes. This paper also justifies the study on bunch graphs which introduced a concept of graphs. where bunch nodes are also allowed. The auto-encoder, a specific type of feedforward neural network generally applied for encoding data in an unsupervised learning methodology to achieve good performance and better-classified data. This kind of network is composed of an encoder and decoder. The encoder compresses the data to an extent or layer, and then from that central layer decoder starts reconstructing the original data. This paper also investigates the dimensionality reduction ability of auto-encoders for character recognition and manipulates the results to accomplish better handling side of auto-encoders. This paper also focuses on the abilities of auto-encoders to reduce noise in data along with dimensionality reduction, trying to interpret the difference between results generated using bunch graph cut techniques. The dataset associated with computing for implementation purposes has been taken from MNIST dataset. Mainly, the two-dimensional plots are used in this paper for comparing results gen crated associated with different parameters that help in recognizing the character as partial and non-partial separabilities.
Keywords:
Bunch graph
Auto-encoders
Dimensionality reduction
Noise reduction
Convolution neural network
Character recognition
Journal
IF:
3
Papers:
1.9W
Citations:
3.2W
View more organizations