TinMan Patterns

TinMan Patterns Features

TinMan Patterns is a data exploration and insight-generating application used on data - usually data in the form of feature vectors - where each feature vector represents a measured identity, circumstance or condition. By providing multiple measurements for a single class and then comparing, registering, matching and analyzing various features/attributes and their relative contribution to accurate classification, an accurate identity signature can be determined. This then can be used in bio-metric systems and non-biometric systems depending on the application.

Comprehensive Visual Exploration of Data Vectors

Import your data as feature vectors and use a broad set of analysis tools to determine which features are most meaningful for defining classes.

Detailed Correlation Analysis of Individual Features

Instantly derive inter and antra class variance across features, along with basic min, max mean statistics. See relative feature level contribution to accurate classification.

Thorough Pattern Matching Analysis with Ranking

See detailed results analysis of euclidian distance pattern matching based on selected set of registered feature vectors. Get ranking and feature level results across vectors.

A Closer Look at TinMan Patterns

Import and Spatially Explore Your Data

Import your csv data as rows of patterns or collections of feature values representing sets of circumstances or attributes of any class. Instantly visualize seperation and grouping within each attribute across classes to determine variance and association.

Easily Perform Matching Tests and Export Results

Determine how many recorded feature vectors for each class are to be used in establishing a registered identity. Then test and instantly perform matching feature by feature vector by vector to determine strength of configured identity. Export results and group by vector or class.

Apply Analytical Tools to Determine Meaningful Features

Apply weighting and exclusion tests on various features to easily see impact on classification results across feature vectors. Rule out meaningless features and add weight to those that help distinguish accurate identity. Apply density maps, filters and range blocks to visually reach conclusions.