Project Goals

Short term goals

• Implement database abstraction library for managing word database.
• Create word database with the following Part-Of-Speech tags:

#	Uzbek	English	Tag
1	От	Noun	NOUN
2	Феьл	Verb	VERB
3	Сифат	Adjective	ADJ
4	Равиш	Adverb	ADV
5	Олмош	Pronoun	PRON
6	Сон	Numeral	NUM
8	Боғловчи	Conjunction	CONJ
9	Юклама	Particle	PART
10	Ундов сўз	Interjection	INTJ
11	Тақлид сўз	Onomatopoeic words	X
12	Модал сўз	Modal words	AUX
13	Кўмакчи	Postposition	ADP

This may not be fully compatible with the Universal Dependencies POS tags.

• Perform statistical analysis on the usage of words, generate a table using PDF books and news feed.
• Implement stemming rules and a table within the database (for the fast reference), table for entity names.
• Implement basic natural language analysis tool that would provide functionalities such tokenizing, parsing, stemming, POS tagging, named entity recognition, etc.

Expected deliverable: The tool tahlih gets an input text in Uzbek language and generates parsed (tokenized, stemmed, POS tagged) output.

Mid term goals

• Manually (Semi-automatically) generate Uzbek Treebank in CoNLL-U format that can be contributed to Universal Dependencies (UD) Framework.
• Feed Uzbek Treebank to SyntaxNet, and perform analysis, training, improving
• Implement initial applications with Uzbek NLU project: Telegram Q&A Bot, Twitter Bot, news summarizer.

Long term goals

• TBD