Algorithms

OVium originates and extends various categories of algorithms to support enterprise methodologies and cloud computing strategies.

  • Search: Crawler agents that seek designated categories of information on internal an external networks, i.e. Droid. Collected documents are processed into inverted indices to support search strategies; ElasticSearch, Solr, CouchDB-Lucene and other lucene based tools.
  • Collection Oriented: Graph and semantic collections are represented as RDF triples in columnar datasets that are quickly and efficiently processed in elastic cloud environments
  • Classification: The employment of clustering and trend prediction algorithms, SVM, LDA and HMM, to characterize categories and identify outliers in data. Outlier items can be used to reprogram search agents to seek new data sources.
  • Natural Language Processing: Development of semantic trees and inverted indexes from documents. NLP products provide inputs and are guided by higher level classification algorithms such as LDA identified topics
  • Security: PGP, public key and private key algorithms for point to point and collaborative security solutions. Basic cyclical and shorter key elliptical algorithms support a wide range of security requirements.
  • Linear Programming: Linear algebra based hyper-polytope and dual constraint solution regions for direct an recursive clustering for classification solutions
  • Non-Linear Programming: Non-linear regions, typically convex, based on higher order splines and dual constraint solution regions for direct an recursive clustering for classification solutions. Other approaches are min-max clustering algorithms that exclusively employ recursive number comparisons rather than standard addition and multiplication arithmetic, eliminates precision loss with alternative classification behavior
  • Finite Field Integer Programming: Finite Galois field algebra and matrix diophantine constraint solution regions for direct an recursive clustering for classification solutions. These solutions use integer processing to eliminate floating point precision errors