COLTEC® | Computer & Language Technology

Technology

Our proprietary linguistic infrastructure is what sets us apart from any competition – our core theory, technologies, algorithms, relations, and database repositories that serve as the foundation for all our applications.

In keeping with our concept that language is a tool used for everyday communication, we have developed an innovative language model to effectively adapt complex and arcane Arabic language concepts for computer use, which is considerably different than the traditional printing methodologies used by our competitors. Because our linguistic model converts language into logical and mathematical models developed specifically for computer use, our applications process language with astonishing speed – far faster than any other programs available today. No other company in the world can make these claims.

We have developed proprietary text and content analysis systems based on our advanced, patent-pending Natural Language Processing (NLP) techniques that are not only unique, but revolutionary in the Arabic language processing industry. Our system guarantees fast, efficient, comprehensive, and above all – intelligent – data analysis and information retrieval capabilities. Our search technologies retrieve information based on concepts, not just keyword matches.

Underlying our technology is a robust Computational Language Theory which, along with our Natural Language Processing Theory, provide the framework upon which our products are built, enabling them to effectively process Arabic content for virtually any language. It’s based on the concept of Language Universals – common patterns shared by all languages. Unlike competitive products that use purely statistical methods and weighting schemes based on machine-learning and frequency, our methodology is linguistics-based, 100% logical, and relies upon our own proprietary linguistic units developed specifically for Arabic. Our model is only supported by statistical methods. Competitive applications rely instead upon word roots which have proven to be less than reliable because they seldom provide the basic meaning for all derivations.

Our methodology instead breaks up the Arabic language into a set of structures and concepts that then form the rules of language usage. Every single Arabic word falls into one or more logical sets and subsets containing features that are automatically assigned to each word – information such as word type, gender, tense, prefixes and suffixes, and much more. Depending on how the word is used and how it needs to be processed, our linguistic tools will reach further into the appropriate sets and subsets to obtain the right data necessary to process the word as accurately as possible.

Because the COLTEC model systematically and comprehensively covers the entire Arabic language with all its rules and details, our applications are capable of fast, efficient, organic growth. As new words and new linguistic uses evolve, they can be easily accommodated under one or more of our existing sets or subsets without reinventing the wheel. And in the rare case that a new set needs to be created, our “open-system” design enables new sets to be added with ease.