| Data Mining - Naive Bayesian V3.2
Introduction
Naive Bayesian is an extremely fast, but less accurate, version of the Bayesian network engine used with classification problems. Naive Bayesian gets its speed from its structural assumption, a condition known as conditional independance amoung the attribute variables. This condition assumes that all of the attribute variables can be considered independant from one another given a known target variable state. The benefit of this assumption is that each of the attribute variables effects can be individually calculated, reducing the whole process to a set of multiplications.
Naive Bayesian networks are commonly used in e-commerce applications, such as web-site personalization, advertisement targeting, and cross-selling because of its raw speed.
Performance
EWA Systems' Naive Bayesian Engine achieves classification speed of over 50,000 inferences per second for the standard Tennis data set even on a 350MHz G3 Processor. This speed is further increased should its multithreading feature be enabled.
Features
- Implemented in 100% Java, but with C-like Performance.
- Unlimited Problem Size (Problem does not have to fit in memory)
- Controlled Multi-Threaded Implementation (Uses only the number of threads specified)
- Uses EWA's Standard Data Manipulation Package
- Features:
- Calculates Priors and Posteriors for All Nodes, not just the Target Node
- Learns Parameters On-The-Fly (Stream-Based Learning) or in Batch-Mode
- Operates with Soft and Hard Evidence, and with Missing Data
To Purchase or For More Information, contact our Sales Team.
|