Pattern Discovery

JMP Genomics Starter | Pattern Discovery

Pattern Discovery

Click on a button corresponding to a pattern discovery process. Refer to the table below for guidance.

Process

Choose this process for...

Hierarchical Clustering

Creating a tree of observation (row) relationships with the option of clustering variables to create a two-way organization, choosing from a variety of methods

K-Means Clustering

Creating optimally separated groups of observations (rows), resulting in groups with similar members

Tip: Consider running Data Standardize before clustering to ensure that the columns are all comparable.

Principal Components Analysis

Examining relationships among many quantitative variables, using orthogonal transformation to reduce potentially correlated variables into uncorrelated variables known as principal components

Plot Intensities

Visualizing row-level intensity measurements for individual samples via Parallel Plots, with the option of placing measurements and plots into groups as defined in an Experimental Design Data Set (EDDS)

Cross Correlation

Computing and testing the significance of all pairwise correlations between two sets of numeric variables, and visualizing these correlations using a Heat Map and Dendrogram

Distance Matrix and Clustering

Computing and plotting distance or dissimilarity measures between observations (rows), and storing these measures in a square matrix output data set that can be used as input for the Multidimensional Scaling process

Multidimensional Scaling

Estimating the coordinates of a set of objects in a space of specified dimensionality (using distance matrix input) and creating a 2-D or 3-D Scatterplot of these coordinates

Tip: Input data for this process can be generated using the Distance Matrix and Clustering process.

Partial Correlation Diagram

Inferring association and potential causal relationships between a set of variables, plotting variables as nodes connected with line segments that vary in appearance based on partial correlations

Tip: A wide variety of plot types and interactive options are available. You are encouraged to explore them all.

See the JMP Genomics Starter main page for other process categories.

Process	Choose this process for...
Hierarchical Clustering	Creating a tree of observation (row) relationships with the option of clustering variables to create a two-way organization, choosing from a variety of methods
K-Means Clustering	Creating optimally separated groups of observations (rows), resulting in groups with similar members Tip: Consider running Data Standardize before clustering to ensure that the columns are all comparable.
Principal Components Analysis	Examining relationships among many quantitative variables, using orthogonal transformation to reduce potentially correlated variables into uncorrelated variables known as principal components
Plot Intensities	Visualizing row-level intensity measurements for individual samples via Parallel Plots, with the option of placing measurements and plots into groups as defined in an Experimental Design Data Set (EDDS)
Cross Correlation	Computing and testing the significance of all pairwise correlations between two sets of numeric variables, and visualizing these correlations using a Heat Map and Dendrogram
Distance Matrix and Clustering	Computing and plotting distance or dissimilarity measures between observations (rows), and storing these measures in a square matrix output data set that can be used as input for the Multidimensional Scaling process
Multidimensional Scaling	Estimating the coordinates of a set of objects in a space of specified dimensionality (using distance matrix input) and creating a 2-D or 3-D Scatterplot of these coordinates Tip: Input data for this process can be generated using the Distance Matrix and Clustering process.
Partial Correlation Diagram	Inferring association and potential causal relationships between a set of variables, plotting variables as nodes connected with line segments that vary in appearance based on partial correlations Tip: A wide variety of plot types and interactive options are available. You are encouraged to explore them all.