Adams-Nafisah/Rule-Based-swahili-POS-tagger

# Rule-Based-swahili-POS-tagger # Rule-Based Part-of-Speech Tagger for Kiswahili ## Overview This project implements a rule-based Part-of-Speech (POS) tagger specifically designed for Kiswahili, an agglutinative language. It leverages morphological analysis (word structure) and contextual analysis (surrounding words) to accurately assign grammatical categories to words. ## Project Aim The objective of this project is to develop a Rule-Based Part-of-Speech Tagger for Kiswahili. Unlike English, Kiswahili is highly agglutinative, meaning a single word often contains the subject, tense, and object (e.g., hawajasafiri = "they have not traveled"). This project aims to identify the grammatical category of words using Morphological Analysis (word structure) and Contextual Analysis (surrounding words). ## Methodology We utilize a two-step algorithm to assign tags: ### Step A: Morphological Tagging (Regex) We apply a prioritized list of Regular Expressions to analyze the internal structure of the word. The order is crucial: * **Closed Classes**: We first check for specific known lists of Pronouns (mimi, wewe), Conjunctions (na, lakini), and Prepositions. * **Orthography**: We check for capitalization to identify Proper Nouns (NNP) and digits for Numbers (CD). * **Affix Analysis**: We analyze prefixes and suffixes common in Kiswahili grammar: * **Prefixes**: Words starting with `ha-` (negative marker) or `ta-` (future tense) are likely verbs. * **Suffixes**: Words ending in `-a` are often Verbs (kula). Words ending in `-o` or `-u` are often Adjectives (mdogo). * **Locatives**: Words ending in `-ni` are treated as Nouns/Locations (nyumbani). ### Step B: Contextual Correction (Brill's Logic) After the initial tagging, we apply "Transformation Rules" to fix errors based on sentence context. **Example**: If a word was tagged as a Noun (NN), but it follows a Preposition (IN), it is likely acting as a Verb (VB) in an infinitive sense (e.g., `kwa kula`). ## Tag …