Predictive Analytics
Predictive Analytics encompasses a variety of techniques from statistics and Data
Mining that analyze current and historical data to make predictions about future events. Our computer scientists have
significant hands-on experience in conceiving, developing and deploying diagnostic and prognostic solutions for Threat Assessment
and Mitigation, Drug Discovery, Integrated Vehicle Health Management (IVHM), Cyber-Security, Wind Turbine monitoring and other Anomaly/Fault
Detection applications.
|
At MAC, we emphasize the three major components of Predictive Analytics:
All aspects of Data Mining / Pattern Recognition / Machine Learning
including detection, classification, regression, clustering and anomaly detection using a variety of powerful
paradigms Data Warehousing for handling massive historical datasets and modern data
streams, such as network traffic, that can generate vast quantities of data in short periods of time Interactive Visualization techniques that enable drill-down and
facilitate hypothesis formation
Data Mining
The Data Exploitation Group’s scientists have extensive experience in designing and implementing Pattern
Recognition / Machine Learning / Data Mining applications. We choose the right paradigm for the problem at hand
based on our expertise. The following table lists some of the techniques we have used in the past to solve
challenging problems in Industry:
|
|
|
Neural Networks
|
Decision Trees and Ensembles
|
Support Vector Machines
|
K Nearest Neighbors
|
Clustering Techniques
|
Bayesian Learning
|
Evolutionary Algorithms
|
Hidden Markov Models
|
Self Organizing Maps (SOMs)
|
Data Warehousing
An extremely important aspect of Predictive Analytics is the collection and maintenance of large
data repositories to enable historical analysis as well as to handle modern monitoring applications capable of generating
terabytes of data in short periods. MAC researchers have experience in designing schema and developing proper strategies
for efficiently storing records using Database platforms such as MySQL, Postgres and Oracle.
Visualization
Choosing the proper techniques for interactively summarizing, filtering and visualizing massive
datasets can make the difference between understanding it and being confused by it. At MAC, we have considerable
understanding of the depth and breadth of visualization techniques and the know-how about what algorithm to use when.
Following is an abbreviated list of techniques that we often use to shed light on datasets:
Frequency Plots
Scatter Plots
Treemaps
Tree Pies
Bubble Plots
Bubble Pies
Bar Charts
Polar Bar Charts
Structured Network Plots
Map Plots
3D Plots and Projection
Pixel-based Views
MAC has the tools and talent to deliver robust Predictive Analytics solutions.
Contact Us 
