Arabic Software  Desktop Publishing  Machine Translation  Document Management  NLP   OCR  ASR  TTS  MultimediA

Proper Nouns Identification System

Concept
AppTek's NameFinder™ is an advanced technology engine that is used to scan text for proper nouns (such as human names) in various languages--even in writing systems that do not use capitalization. NameFinder™ can also accurately identify the ethno-linguistic origins of a person's name.
The system recognizes transliterates between the Roman (Latin) alphabet and other writing systems, and can identify transliterated names despite discrepancies, ambiguities, or simple misspellings. The system is designed to assist customers with multilingual name lexicons. Proper nouns may label locations, persons, currencies, measurements, first or last names, etc.

The system holds a large database of proper nouns, which may be associated with other nouns and types of events, regardless of their proximity in a given text; it also identifies nouns in free text, using surface parsing.

Components
AppTek NameFinder™ is comprised of the following components:
Proper noun normalization
Analysis and transliteration
Universal name identifier (UNI)
Rich proper nouns lexicon in several languages.
Proper noun detection in free -text.
NameFinder™ operates on all Windows Platforms. It has dynamic DLL and API calls for integration with other Systems. The system can be deployed with various databases including B-tree, ODBC, SQL Server, Oracle, etc.

Functions
NameFinder™ uses local and remote context and statistical information to determine if a word is a potential proper noun. It looks for words that are flagged as “can be a name” in the lexicon, and context clues (such as “Mr.”, “doctor”.) to help find other names. The remote context incorporates verbs and words in general that are associated (collocation) with proper nouns. These are determined statistically by using the running text corpora. If a word, or words, are flagged as “can be a name” in the lexicon and context information enforces the proper name detection, the word is tagged as being a proper name. Words that do not meet both criteria are tagged as potential proper nouns.

NameFinder™ Bridge the Language Gap

AppTek NameFinder™ provides the following capabilities:
Utilizes special linguistic algorithms
Detect, analyze and transliterate proper nouns
Detects Proper nouns in free-format text
Validate name entries and transliteration
Identify possible similarity or match between various spelling of proper noun
Identify specific ethno linguistic name groups
Handles name spelling ambiguity Roman (Latin) vs. non-Latin
Check against standard (base-form) transliteration entry in the Lexicon
Link discrepant entry with standard lexical entry
Produce various reports on discrepant, rejected, and valid entries.
Provide system administrator with maintenance utility for Lexicon
Integrates with Machine Translation for Name-Context analysis
Supports multiple languages and scripts such as English/Latin, Cyrillic, Arabic, Farsi, etc.

Please contact George N. Hallak for pricing and customized solutions.