// THESIS GUIDANCE PORTAL · Computer Science

AI Chatbot for Nepali Government Service Guidance

Level: BachelorDifficulty: Beginner★ Popular Choice

1. Introduction & Problem Statement

Overview: NLP chatbot answering citizen questions about citizenship, passport, and ward-office procedures in Nepali.

Background Context (Nepal): Citizens in Nepal often face confusion regarding required documentation and fee structures for ward office services. Conversational AI in Devanagari provides 24/7 procedural guidance.

2. Research Objectives

›Scrape and curate structured FAQ dataset of government service procedures (Citizenship, Passport, Vital Registration)
›Build bilingual (Nepali Devanagari and Romanized Nepali) intent recognition system using RAG / Intent Transformers
›Develop conversational web interface with voice input capability
›Integrate official ward office fee tables and document checklists
›Evaluate chatbot accuracy, intent precision, and response generation time

3. Proposed Methodology

Data collection from official Ministry of Home Affairs and Kathmandu Metropolitan City administrative portals
Natural Language Processing using Devanagari text normalization, stop-word removal, and embedding generation
Retrieval-Augmented Generation (RAG) architecture using LangChain, FAISS vector store, and open LLMs
Frontend widget interface built with React/Next.js and Web Speech API
Usability and intent evaluation across 200 sample citizen queries

$ Worked Example / Sample Scenario

Sample Scenario: A student asks in Romanized Nepali 'Nagarikta banauna ke ke kaagajat chahincha?'. The RAG chatbot identifies the intent, retrieves the official MoHA checklist, and presents a clear bulleted list of 4 required documents in Devanagari.

4. Thesis Chapter-by-Chapter Outline

Chapter 1: IntroductionTU/KU standard

Background, problem statement, research questions, objectives, scope, limitations, and significance of the study

Chapter 2: Literature ReviewTU/KU standard

Theoretical framework, conceptual models, previous empirical studies in Nepal and developing nations, Devanagari text tokenization and Retrieval-Augmented Generation (RAG) design, and gap analysis

Chapter 3: Research MethodologyTU/KU standard

Research design, population/sampling framework, data collection instruments, analytical tools, and ethical considerations

Chapter 4: Data Analysis & ResultsTU/KU standard

Empirical findings, statistical testing, model estimations, intent classification confusion matrix and RAG accuracy, and detailed discussion

Chapter 5: Conclusion & RecommendationsTU/KU standard

Summary of key findings, theoretical contributions, policy recommendations, and future research directions

5. Recommended Tools & Technologies

To implement the practical, technical, or analytical portions of this thesis topic, the following software tools, libraries, or APIs are recommended:

PythonLangChain / LlamaIndexFAISS / QdrantNext.jsOpenAI / Local LLMTailwind CSS

6. Core References & Academic Sources

[1]Government of Nepal — Ministry of Home Affairs (MoHA) Citizen Charter Guidelines
[2]Kathmandu Metropolitan City — Ward Office Service Portals & Citizen Charters
[3]NepJOL / IEEE Xplore — Devanagari Natural Language Processing and Chatbot systems

7. Frequently Asked Questions (FAQs)

Q: How do you handle both Devanagari and Romanized Nepali (e.g., 'Nagarikta')?

You train or fine-tune a transliteration preprocessing step or multi-lingual embedding model that maps Romanized phrases to Devanagari equivalents.

Q: What is RAG and why is it useful for government service chatbots?

Retrieval-Augmented Generation (RAG) grounds the AI response strictly on official government documents, preventing hallucinated or incorrect information.

Q: Can this run using free open-source models?

Yes, using open models like LLaMA-3 or Mistral via Ollama combined with local vector stores like FAISS ensures zero API cost.

Q: Where can I get official document checklists for Nepal services?

Official Citizen Charters (Nagarik Charter) displayed on ward office websites and Ministry of Home Affairs portals serve as verified source data.