Fig 2 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Download
View publication
Self-Organizing Map (SOM) Neighbor Weight Distances depicting SOM Topology. Horizontal axis indicates the number of neurons.

Self-Organizing Map (SOM) Neighbor Weight Distances depicting SOM Topology. Horizontal axis indicates the number of neurons.

Source publication
Fig. 1 (a) Experimental setup for reconstruction of Zernike...
Fig. 2 Self-Organizing Map (SOM) Neighbor Weight Distances depicting...
Fig. 3 SOM Sample Hits depicting the number of input vectors linked to...
Fig. 5 SOM Weight Planes displaying the locations of data points and...
+2 Fig. 6 Dendrogam for clusters.
Comparison between neural network clustering, hierarchical clustering and k-means clustering: Applications using fluidic lenses
Preprint
Full-text available
  • Mar 2025
A comparison between neural network clustering (NNC), hierarchical clustering (HC) and K-means clustering (KMC) is performed to evaluate the computational superiority of these three machine learning (ML) techniques for organizing large datasets into clusters. For NNC, a self-organizing map (SOM) training was applied to a collection of wavefront sen...
Cite
Download full-text

Contexts in source publication

Context 1
... an input of 15 Zernike variables, a neural network (NN) consisting of 10 neurons, as shown in Fig. 1(b), was trained until it reached the stopping criteria of 200 epochs of the ML Batch algrorithm. The dimensions of the Self-Organizing-Map (SOM) corresponds to the number of neurons illustrated on the horizontal axis in Fig. 2. In Fig. 2, data comprising 8 predictor variables is input into an SOM network, allowing neurons to layer and group based on shared characteristics. Ultimately, the SOM organizes the related neurons into a 100-grid map. ...
Context 2
... an input of 15 Zernike variables, a neural network (NN) consisting of 10 neurons, as shown in Fig. 1(b), was trained until it reached the stopping criteria of 200 epochs of the ML Batch algrorithm. The dimensions of the Self-Organizing-Map (SOM) corresponds to the number of neurons illustrated on the horizontal axis in Fig. 2. In Fig. 2, data comprising 8 predictor variables is input into an SOM network, allowing neurons to layer and group based on shared characteristics. Ultimately, the SOM organizes the related neurons into a 100-grid map. ...
Context 3
... represent all weights within an 8-dimensional input space-consisting of 8 different fluidic volumen samples for each of the 15 input Zernike variables-we create a plot of SOM neighbor weight distances, illustrated in Fig. 2. This plot is characterized by a hexagonal grid known as SOM topology, where each blue hexagon corresponds to a neuron. Throughout the training phase, the weight vector linked to each neuron adjusts to become the center of a cluster of input vectors. The attributes of the plot are clarified using color-coding: blue hexagons signify the ...
Context 4
... the distribution and consolidation of Z 1 − Z 15 Zernike components. The visualization of sample hits suggests that the 16 sampled locations can be categorized into eight distinct clusters based on data conformity (Fig. 3). In addition to this, the dissimilarity among the 16 vectors was assessed through Neighbour Weight Distances as shown in Fig. 2. Approximately 8 clusters can be identified according to high correlation, indicated by yellow areas surrounded by dark red or black patches in the maps. To identify highly correlated Zernike variables, a weight plane visualization (Fig. 4) was generated to display the weights associated with each parameter for individual neurons. ...