Tuesday, January 7, 2020

Solving The Physics Of The Problem - 1393 Words

As the name suggests, there are no basic guidelines for these algorithms, hence it is unsupervised. These algorithms can be used to discover various pattern, divide the data into various clusters, reducing the dimensionality of the dataset for viewing, which may help researchers in better understanding of the physics of the problem. Here, an expert needs to be careful while choosing a certain algorithm and associated parameters for a specific case. Additionally, an expert needs to be very careful while interpreting the findings from these algorithms. One must use the technical aspects regarding the basic physics of the problem so that their results are meaningful and for it to be accepted by the materials research specialists for†¦show more content†¦The final result is a tree like structure referred as Dendrogram, which shows the way the clusters are related. User can specify a distance or number of clusters to view the dataset in disjoint groups. In this way, the user can get rid of a cluster that does not serve any purpose as per his expertise. In this case, we used MVA (Multivariate data analysis) node in optimization package: modeFRONTIER (ESTECO, 2015) and other statistical software IBM SPSS (IBMSPSS, 2015) for HCA analysis. Clusters are classified by following measures (ESTECO, 2015) 1. Internal similarity (ISim): It reflects the compactness of the k-th cluster. It must be higher. 2. External similarity (ESim): It reflects the uniqueness of the k - th cluster. It must be lower. 3. Descriptive variables: are the most significant variables that help in identifying cluster elements that are similar to one another. 4. Discriminating variables: are the most significant variables that help in identifying cluster elements that are dissimilar to other clusters. HCA analysis can be used to cross check the findings of SVR analysis mentioned above in the text. 4.3.2 Principal Component Analysis (PCA) Principal component analysis can be classified as an unsupervised learning machine-learning algorithm [Mueller et~al., 2015]. It was performed in order to determine correlations

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.