README.md

---
language: lo
language_name: Lao
language_family: taikadai_southwestern
tags:
  - wikilangs
  - nlp
  - tokenizer
  - embeddings
  - n-gram
  - markov
  - wikipedia
  - feature-extraction
  - sentence-similarity
  - tokenization
  - n-grams
  - markov-chain
  - text-mining
  - fasttext
  - babelvec
  - vocabulous
  - vocabulary
  - monolingual
  - family-taikadai_southwestern
license: mit
library_name: wikilangs
pipeline_tag: text-generation
datasets:
  - omarkamali/wikipedia-monthly
dataset_info:
  name: wikipedia-monthly
  description: Monthly snapshots of Wikipedia articles across 300+ languages
metrics:
  - name: best_compression_ratio
    type: compression
    value: 4.276
  - name: best_isotropy
    type: isotropy
    value: 0.7907
  - name: vocabulary_size
    type: vocab
    value: 0
generated: 2026-01-10
---

# Lao - Wikilangs Models
## Comprehensive Research Report & Full Ablation Study

This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Lao** Wikipedia data.
We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and word embeddings.

## 📋 Repository Contents

### Models & Assets

- Tokenizers (8k, 16k, 32k, 64k)
- N-gram models (2, 3, 4, 5-gram)
- Markov chains (context of 1, 2, 3, 4 and 5)
- Subword N-gram and Markov chains
- Embeddings in various sizes and dimensions (aligned and unaligned)
- Language Vocabulary
- Language Statistics

![Performance Dashboard](visualizations/performance_dashboard.png)

### Analysis and Evaluation

- [1. Tokenizer Evaluation](#1-tokenizer-evaluation)
- [2. N-gram Model Evaluation](#2-n-gram-model-evaluation)
- [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
- [4. Vocabulary Analysis](#4-vocabulary-analysis)
- [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
- [6. Morphological Analysis (Experimental)](#6--morphological-analysis-experimental)
- [7. Summary & Recommendations](#7-summary--recommendations)
- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
- [Visualizations Index](#visualizations-index)

---
## 1. Tokenizer Evaluation

![Tokenizer Compression](visualizations/tokenizer_compression.png)

![Tokenizer Fertility](visualizations/tokenizer_fertility.png)

![Tokenizer OOV](visualizations/tokenizer_oov.png)

![Total Tokens](visualizations/tokenizer_total_tokens.png)

### Results

| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
|------------|-------------|---------------|----------|--------------|
| **8k** | 3.254x | 3.29 | 0.1618% | 425,150 |
| **16k** | 3.649x | 3.69 | 0.1814% | 379,183 |
| **32k** | 4.002x | 4.05 | 0.1990% | 345,735 |
| **64k** | 4.276x 🏆 | 4.32 | 0.2126% | 323,587 |

### Tokenization Examples

Below are sample sentences tokenized with each vocabulary size:

**Sample 1:** `.gf ແມ່ນໂດເມນອິນເຕີເນັດລະດັບສູງສຸດຕາມລະຫັດປະເທດ (ccTLD) ສໍາລັບປະເທດເຟນົເກີຍ`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁. g f ▁ແມ່ນ ໂດເມນອິນເຕີເນັດ ລະດັບສູງສຸດ ຕາມລະຫັດປະເທດ ▁( cctld ) ... (+6 more)` | 16 |
| 16k | `▁. g f ▁ແມ່ນ ໂດເມນອິນເຕີເນັດ ລະດັບສູງສຸດ ຕາມລະຫັດປະເທດ ▁( cctld ) ... (+6 more)` | 16 |
| 32k | `▁. g f ▁ແມ່ນ ໂດເມນອິນເຕີເນັດ ລະດັບສູງສຸດ ຕາມລະຫັດປະເທດ ▁( cctld ) ... (+5 more)` | 15 |
| 64k | `▁. gf ▁ແມ່ນ ໂດເມນອິນເຕີເນັດ ລະດັບສູງສຸດ ຕາມລະຫັດປະເທດ ▁( cctld ) ▁ສໍາລັບປະເທດ ... (+2 more)` | 12 |

**Sample 2:** `6 ສິງຫາ ເປນວັນທີ່ 218 ຊອງປີຕາມປະຕິທິນສຸລິຍະຄະຕິ ເຫດການສຳຄັນ ວັນເກີດ ວັນເສຍຊີວິດ`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁ 6 ▁ສິງຫາ ▁ເປນວັນທີ່ ▁ 2 1 8 ▁ຊອງປີຕາມປະຕິທິນ ສຸລິຍະຄະຕິ ... (+3 more)` | 13 |
| 16k | `▁ 6 ▁ສິງຫາ ▁ເປນວັນທີ່ ▁ 2 1 8 ▁ຊອງປີຕາມປະຕິທິນ ສຸລິຍະຄະຕິ ... (+3 more)` | 13 |
| 32k | `▁ 6 ▁ສິງຫາ ▁ເປນວັນທີ່ ▁ 2 1 8 ▁ຊອງປີຕາມປະຕິທິນ ສຸລິຍະຄະຕິ ... (+3 more)` | 13 |
| 64k | `▁ 6 ▁ສິງຫາ ▁ເປນວັນທີ່ ▁ 2 1 8 ▁ຊອງປີຕາມປະຕິທິນ ສຸລິຍະຄະຕິ ... (+3 more)` | 13 |

**Sample 3:** `ສຸດທິດາ ອະນຸສິນ (ຊື່ຫຼິ້ນ:ມິມີ່) ເປັນນາງງາມ,ນາງແບບຊາວລາວ ແລະເປັນຕົວແທນຈາກແຂວງສະຫ...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁ສຸດ ທິດາ ▁ອະນຸ ສິນ ▁( ຊື່ຫຼິ້ນ : ມິ ມີ ່ ... (+18 more)` | 28 |
| 16k | `▁ສຸດ ທິດາ ▁ອະນຸ ສິນ ▁( ຊື່ຫຼິ້ນ : ມິ ມີ່ ) ... (+14 more)` | 24 |
| 32k | `▁ສຸດທິດາ ▁ອະນຸສິນ ▁( ຊື່ຫຼິ້ນ : ມິ ມີ່ ) ▁ເປັນນາງງາມ , ... (+12 more)` | 22 |
| 64k | `▁ສຸດທິດາ ▁ອະນຸສິນ ▁( ຊື່ຫຼິ້ນ : ມິ ມີ່ ) ▁ເປັນນາງງາມ , ... (+10 more)` | 20 |


### Key Findings

- **Best Compression:** 64k achieves 4.276x compression
- **Lowest UNK Rate:** 8k with 0.1618% unknown tokens
- **Trade-off:** Larger vocabularies improve compression but increase model size
- **Recommendation:** 32k vocabulary provides optimal balance for production use

---
## 2. N-gram Model Evaluation

![N-gram Perplexity](visualizations/ngram_perplexity.png)

![N-gram Unique](visualizations/ngram_unique.png)

![N-gram Coverage](visualizations/ngram_coverage.png)

### Results

| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
|--------|---------|------------|---------|----------------|------------------|-------------------|
| **2-gram** | Word | 3,552 | 11.79 | 7,139 | 20.1% | 49.9% |
| **2-gram** | Subword | 2,184 🏆 | 11.09 | 20,562 | 30.0% | 71.8% |
| **3-gram** | Word | 5,061 | 12.31 | 8,179 | 15.3% | 43.3% |
| **3-gram** | Subword | 18,283 | 14.16 | 110,536 | 10.7% | 33.7% |
| **4-gram** | Word | 15,243 | 13.90 | 20,416 | 8.0% | 22.3% |
| **4-gram** | Subword | 79,610 | 16.28 | 344,446 | 5.8% | 18.2% |
| **5-gram** | Word | 12,635 | 13.63 | 15,847 | 8.0% | 21.2% |
| **5-gram** | Subword | 160,221 | 17.29 | 489,758 | 3.6% | 12.7% |

### Top 5 N-grams by Size

**2-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `ຄ ສ` | 3,689 |
| 2 | `พ ศ` | 1,013 |
| 3 | `ພ ສ` | 932 |
| 4 | `of the` | 838 |
| 5 | `ປະ ເທດ` | 375 |

**3-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `ສ ຄ ສ` | 245 |
| 2 | `ສາ ທາ ລະ` | 193 |
| 3 | `ທາ ລະ ນະ` | 186 |
| 4 | `ພ ສ ຄ` | 185 |
| 5 | `ຊອງປີຕາມປະຕິທິນສຸລິຍະຄະຕິ ເຫດການສຳຄັນ ວັນເກີດ` | 169 |

**4-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `ພ ສ ຄ ສ` | 185 |
| 2 | `ສາ ທາ ລະ ນະ` | 183 |
| 3 | `ທາ ລະ ນະ ລັດ` | 164 |
| 4 | `ຊອງປີຕາມປະຕິທິນສຸລິຍະຄະຕິ ເຫດການສຳຄັນ ວັນເກີດ ວັນເສຍຊີວິດ` | 164 |
| 5 | `ສະບັບຄົ້ນຄວ້າ ສົມມະນາ ການພິມ ສປປ` | 139 |

**5-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `ສາ ທາ ລະ ນະ ລັດ` | 163 |
| 2 | `ສະບັບຄົ້ນຄວ້າ ສົມມະນາ ການພິມ ສປປ ລາວ` | 139 |
| 3 | `ສ ສະບັບຄົ້ນຄວ້າ ສົມມະນາ ການພິມ ສປປ` | 139 |
| 4 | `ຄ ສ ສະບັບຄົ້ນຄວ້າ ສົມມະນາ ການພິມ` | 139 |
| 5 | `ພ ສ ຄ ສ ສະບັບຄົ້ນຄວ້າ` | 139 |

**2-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `າ ນ` | 59,706 |
| 2 | `ອ ງ` | 44,102 |
| 3 | `ນ _` | 39,271 |
| 4 | `ກ າ` | 37,742 |
| 5 | `ລ ະ` | 34,362 |

**3-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `ກ າ ນ` | 28,652 |
| 2 | `ແ ລ ະ` | 21,130 |
| 3 | `ຂ ອ ງ` | 19,934 |
| 4 | `ເ ປັ ນ` | 16,912 |
| 5 | `_ ແ ລ` | 15,559 |

**4-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `_ ແ ລ ະ` | 15,503 |
| 2 | `ແ ລ ະ _` | 10,398 |
| 3 | `ຄ ວ າ ມ` | 6,185 |
| 4 | `ະ ເ ທ ດ` | 5,750 |
| 5 | `ປ ະ ເ ທ` | 5,748 |

**5-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `_ ແ ລ ະ _` | 9,660 |
| 2 | `ປ ະ ເ ທ ດ` | 5,720 |
| 3 | `_ t h e _` | 3,835 |
| 4 | `ຄ . ສ . _` | 3,241 |
| 5 | `_ ຄ . ສ .` | 2,982 |


### Key Findings

- **Best Perplexity:** 2-gram (subword) with 2,184
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
- **Coverage:** Top-1000 patterns cover ~13% of corpus
- **Recommendation:** 4-gram or 5-gram for best predictive performance

---
## 3. Markov Chain Evaluation

![Markov Entropy](visualizations/markov_entropy.png)

![Markov Contexts](visualizations/markov_contexts.png)

![Markov Branching](visualizations/markov_branching.png)

### Results

| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
|---------|---------|-------------|------------|------------------|-----------------|----------------|
| **1** | Word | 0.3040 | 1.235 | 2.14 | 178,626 | 69.6% |
| **1** | Subword | 1.0554 | 2.078 | 12.72 | 4,475 | 0.0% |
| **2** | Word | 0.1004 | 1.072 | 1.19 | 378,257 | 90.0% |
| **2** | Subword | 0.7301 | 1.659 | 5.16 | 56,905 | 27.0% |
| **3** | Word | 0.0354 | 1.025 | 1.06 | 447,523 | 96.5% |
| **3** | Subword | 0.5034 | 1.418 | 2.77 | 293,770 | 49.7% |
| **4** | Word | 0.0171 🏆 | 1.012 | 1.02 | 468,823 | 98.3% |
| **4** | Subword | 0.3413 | 1.267 | 1.85 | 811,856 | 65.9% |

### Generated Text Samples (Word-based)

Below are text samples generated from each word-based Markov chain model:

**Context Size 1:**

1. `ແລະ ຄວາມຊົ່ວ ຫລື ໃນຊາຕເທົ່ານັ້ນບໍ່ແມ່ນເອົາສຫວັນເມື່ອຕາຍແລ້ວ ແລະ ລາງ ເຊັ່ນດຽວກັນກັບຫ້ອງໂຖງສະຖານີທີ່ສາ...`
2. `ສ ເຖິງ 12 07 ມັງກອນ ເປັນ ລູກ ດ້ວຍ ເຖົ້າ ພໍກະ ເທີນ ເຖົ້າ ກະ ສານ ຂອງ ກະຊວງກະສິກຳແລະປ່າໄມ້`
3. `the summation takes ເພື່ອປ້ອງກັນບໍ່ໃຫ້ສະ ກຸນເງິນເອີໂຣ ລົ້ມສະຫລາຍ ໃນປີ ນະຄອນ ທີ່ມີຕັ້ງແຕ່ປະມານສະຕະວັດ...`

**Context Size 2:**

1. `ຄ ສ ເມຖຸນາ ຄ ສ ຫາ ໂດຍມີກຸງສີອະຍຸດທະຍາເປັນນະຄອນຫຼວງ ອານາຈັກອະຍຸດທະຍານັບວ່າຈະເລີນຮຸ່ງເຮືອງຮັ່ງມີທີ່ສຸດ...`
2. `พ ศ ราชกิจจานุเบกษา สืบค้นวันที่ 25 มิถุนายน ຕໍ່ມາວັນທີ 17 ພະຈິກ ທີ່ເມືອງກາມາວ ແຂວງກາມາວ ການສຶກສາ ຈົ...`
3. `of the 20th century fox ອ້າງອີງ ລິ້ງຄ໌ພາຍນອກ sukhoi su 27 ທີ່ສຸດ ມັກຈະ ມີແມງວັນ ຄະ ດີກວ່າ ອາກາດ ແຕ່`

**Context Size 3:**

1. `ສ ຄ ສ ສະບັບຄົ້ນຄວ້າ ສົມມະນາ ການພິມ ສປປ ລາວ ເປັນລະດູອັນມີອາກາດໜາວທີ່ສຸດໃນປີ ລະດູໜາວໃນເຂດອົບອຸ່ນແລະເຂດ...`
2. `ສາ ທາ ລະ ນະ ລັດ ເຢ ເມນ 45 ໝູ່ເກາະເຄ ແມນ 46 ການາດາ 47 ສາ ທາ ລະ ນະ ຂອງ`
3. `ທາ ລະ ນະ ລັດ ປະ ຊາ ຊົນ ຈີນ 48 ຈີນ ໄທ ເປ ໄຕ້ ຫວັນ 49 ອາ ນາ ເຂດ`

**Context Size 4:**

1. `ສາ ທາ ລະ ນະ ລັດ ປະ ຊາ ທິ ປະ ໄຕ ແລະ ການ ປົກ ຄອງ ໂດຍ ກົງ ຂອງ ມົງ ກຸດ`
2. `ພ ສ ຄ ສ ຕະຫຼອດໄລຍະເວລາທີລາຍການລະເບີດເຖິດເທິງອອກອາກາດຈົນຮອດປະຈຸບັນ ຊ່ວງຊິດຄອມລະເບີດເຖິດເທິງໃນຍຸກຕໍ່ໆມ...`
3. `ທາ ລະ ນະ ລັດ ບູ ຣຸນ ດີ 97 ສາ ທາ ລະ ນະ ຂອງ ອັ ຟ ກາ ນິ ສ ຖານ`


### Generated Text Samples (Subword-based)

Below are text samples generated from each subword-based Markov chain model:

**Context Size 1:**

1. `_vens_ໄຊາ​ປະຊ_sh`
2. `ານ_ຂ້ອຍ​ສ້າມວົງກາວລ`
3. `ນ​ກລັດທຶນ​ອົງປີ_ລະວົງ`

**Context Size 2:**

1. `ານງ​ໂດຍໄປນີ້_ເພື່ອເຈຍ`
2. `ອງຢ້ຽມເຕະແຂວງ_3100`
3. `ນ_ໆ_ເປັນເມືອງຍ່ອຍຮັບປ`

**Context Size 3:**

1. `ການເລື່ອງທີ່ມີຊີວິດແຕ່ພາເວ`
2. `ແລະ_ກຸງຣາຊະຖາບັນເທົາ_`
3. `ຂອງກ້ອນມົກກະພັດຍີ່ປຸ່ນຈະຕຸ`

**Context Size 4:**

1. `_ແລະ_ຕຸ໊ກກີ້_ຊິງຮ້ອຍຊິງລ້ານ`
2. `ແລະ_ປັບປຸງເພື່ອພະສົບແລ້ວ,`
3. `ຄວາມຮັກສາພະອົງເຈົ້າພະຍາ`


### Key Findings

- **Best Predictability:** Context-4 (word) with 98.3% predictability
- **Branching Factor:** Decreases with context size (more deterministic)
- **Memory Trade-off:** Larger contexts require more storage (811,856 contexts)
- **Recommendation:** Context-3 or Context-4 for text generation

---
## 4. Vocabulary Analysis

![Zipf's Law](visualizations/zipf_law.png)

![Top Words](visualizations/top20_words.png)

![Coverage Curve](visualizations/vocab_coverage.png)

### Statistics

| Metric | Value |
|--------|-------|
| Vocabulary Size | 36,378 |
| Total Tokens | 382,077 |
| Mean Frequency | 10.50 |
| Median Frequency | 3 |
| Frequency Std Dev | 92.90 |

### Most Common Words

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | ແລະ | 10,966 |
| 2 | ສ | 5,253 |
| 3 | the | 4,001 |
| 4 | ຄ | 3,718 |
| 5 | 1 | 3,286 |
| 6 | of | 3,039 |
| 7 | ຂອງ | 2,967 |
| 8 | ໃນ | 2,847 |
| 9 | 2 | 2,607 |
| 10 | ການ | 2,393 |

### Least Common Words (from vocabulary)

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | ປະຕິເສດຄວາມຮັບຜິດຊອບແລະຖືກຕໍາຫຼວດຈັບຕົວເພື່ອສອບຖາມໃນວັນທີ | 2 |
| 2 | ຜູ້ເຄາະຮ້າຍ | 2 |
| 3 | ຢ່າງນ້ອຍ | 2 |
| 4 | ຍ້ອນການສົງໄສວ່າ | 2 |
| 5 | ດື່ມເຄື່ອງດື່ມທີ່ມີສານ | 2 |
| 6 | ແລະອີກ | 2 |
| 7 | ເຈັບປ່ວຍຈາກການບໍລິໂພກເຫຼົ້າປົນເປື້ອນນັ້ນ | 2 |
| 8 | ໂຣມາຍາ | 2 |
| 9 | ເຈເຣັດ | 2 |
| 10 | ລະດູຮ້ອນໃນເມືອງແມ່ນຮ້ອນແລະຊຸ່ມຊື່ນ | 2 |

### Zipf's Law Analysis

| Metric | Value |
|--------|-------|
| Zipf Coefficient | 0.9640 |
| R² (Goodness of Fit) | 0.996857 |
| Adherence Quality | **excellent** |

### Coverage Analysis

| Top N Words | Coverage |
|-------------|----------|
| Top 100 | 28.8% |
| Top 1,000 | 54.5% |
| Top 5,000 | 74.0% |
| Top 10,000 | 82.5% |

### Key Findings

- **Zipf Compliance:** R²=0.9969 indicates excellent adherence to Zipf's law
- **High Frequency Dominance:** Top 100 words cover 28.8% of corpus
- **Long Tail:** 26,378 words needed for remaining 17.5% coverage

---
## 5. Word Embeddings Evaluation

![Embedding Isotropy](visualizations/embedding_isotropy.png)

![Similarity Matrix](visualizations/embedding_similarity.png)

![t-SNE Words](visualizations/tsne_words.png)

![t-SNE Sentences](visualizations/tsne_sentences.png)


### 5.1 Cross-Lingual Alignment

![Alignment Quality](visualizations/embedding_alignment_quality.png)

![Multilingual t-SNE](visualizations/embedding_tsne_multilingual.png)


### 5.2 Model Comparison

| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
|-------|-----------|----------|------------------|---------------|----------------|
| **mono_32d** | 32 | 0.7907 🏆 | 0.3413 | N/A | N/A |
| **mono_64d** | 64 | 0.4812 | 0.3220 | N/A | N/A |
| **mono_128d** | 128 | 0.1381 | 0.3128 | N/A | N/A |
| **aligned_32d** | 32 | 0.7907 | 0.3454 | 0.0400 | 0.2780 |
| **aligned_64d** | 64 | 0.4812 | 0.3164 | 0.0760 | 0.3560 |
| **aligned_128d** | 128 | 0.1381 | 0.3154 | 0.1140 | 0.4340 |

### Key Findings

- **Best Isotropy:** mono_32d with 0.7907 (more uniform distribution)
- **Semantic Density:** Average pairwise similarity of 0.3256. Lower values indicate better semantic separation.
- **Alignment Quality:** Aligned models achieve up to 11.4% R@1 in cross-lingual retrieval.
- **Recommendation:** 128d aligned for best cross-lingual performance

---
## 6.  Morphological Analysis (Experimental)

This section presents an automated morphological analysis derived from the statistical divergence between word-level and subword-level models. By analyzing where subword predictability spikes and where word-level coverage fails, we can infer linguistic structures without supervised data.

### 6.1 Productivity & Complexity

| Metric | Value | Interpretation | Recommendation |
|--------|-------|----------------|----------------|
| Productivity Index | **5.000** | High morphological productivity | Reliable analysis |
| Idiomaticity Gap | **0.675** | High formulaic/idiomatic content | - |

### 6.2 Affix Inventory (Productive Units)

These are the most productive prefixes and suffixes identified by sampling the vocabulary for global substitutability patterns. A unit is considered an affix if stripping it leaves a valid stem that appears in other contexts.

#### Productive Prefixes
| Prefix | Examples |
|--------|----------|
| `-ເ` | ເມືອງນາກາຍ, ເສ, ເບສດີໄຊເນີ |
| `-ສ` | ສິງເສນ, ສະຖານີສາມຢ່ານ, ສົມເດັດພະສີພັດຊະລິນທາບໍລົມລາຊິນີນາດ |
| `-s` | skants, stage, scandinavia |
| `-ອ` | ອິງຟ້າ, ອະນຸສາວະລີ, ອະດຸນຍະສັກອັກຄະນະເຣດສະຣາທິບໍດີ |
| `-a` | ashgabat, at, alison |
| `-ການ` | ການສຸ່ມກາຊາ, ການທີ່ຊາວບ້ານຍັງເຊື່ອຖືສິ່ງດັ່ງກ່າວແມ່ນບໍ່ໄດ້ຖື, ການປະຕິວັດເດືອນຕຸລາ |
| `-ກ` | ກຽດຕິຍົດຄົນທີວີ, ການສຸ່ມກາຊາ, ການທີ່ຊາວບ້ານຍັງເຊື່ອຖືສິ່ງດັ່ງກ່າວແມ່ນບໍ່ໄດ້ຖື |
| `-m` | men, martini, mans |

#### Productive Suffixes
| Suffix | Examples |
|--------|----------|
| `-ນ` | ສິງເສນ, ປະຍູນພັນ, ໄລຍະທາງປະມານ |
| `-ງ` | ນະຄອນຫລວງ, ຕະວັນພົງ, ພຽງແມ່ນ້ຳຂອງ |
| `-ດ` | ຕົ້ນກຳເນີດ, ບ້ານຄຳສະຫວາດ, ແຄ່ຄົນໂທຜິດ |
| `-າ` | ແຄນນາດາ, ອິງຟ້າ, ດາ |
| `-s` | skants, mans, holdings |
| `-ານ` | ໄລຍະທາງປະມານ, ເຄີດິສຖານ, ສະຖານີສາມຢ່ານ |
| `-ກ` | ຄາລສະເບີກ, ນຄນປລັນໂລກ, ລັອກ |
| `-e` | treble, stage, unicode |

### 6.3 Bound Stems (Lexical Roots)

Bound stems are high-frequency subword units that are semantically cohesive but rarely appear as standalone words. These often correspond to the 'core' of a word that requires inflection or derivation to be valid.

| Stem | Cohesion | Substitutability | Examples |
|------|----------|------------------|----------|
| `tion` | 2.20x | 21 contexts | nation, notion, motion |
| `atio` | 2.22x | 15 contexts | ratio, nation, nations |
| `ະເທດ` | 1.87x | 12 contexts | ຮະເທດ, ປະເທດ, ປີປະເທດ |
| `ະການ` | 1.71x | 13 contexts | ມະການ, ປະການ, ແລະການ |
| `ປະເທ` | 1.78x | 10 contexts | ປະເທດ, ປີປະເທດ, ປະເທດໄທ |
| `ນການ` | 1.94x | 8 contexts | ໃນການ, ບວນການ, ແຜນການ |
| `ປະກອ` | 1.65x | 12 contexts | ປະກອບ, ປະກອນ, ຄຳປະກອບ |
| `ານປະ` | 1.99x | 7 contexts | ການປະກອບ, ການປະສູດ, ການປະກັນ |
| `ສະຖາ` | 2.02x | 6 contexts | ສະຖານ, ສະຖານະ, ສະຖານີ |
| `ງປະເ` | 2.17x | 5 contexts | ຂອງປະເທດ, ແດງປະເສີດ, ແຂວງປະເວດ |
| `ງຈາກ` | 1.83x | 7 contexts | ຮອງຈາກ, ບາງຈາກ, ຫຼັງຈາກ |
| `ທະຍາ` | 2.05x | 5 contexts | ວິທະຍາ, ພັດທະຍາ, ວິທະຍາໄລ |

### 6.4 Affix Compatibility (Co-occurrence)

This table shows which prefixes and suffixes most frequently co-occur on the same stems, revealing the 'stacking' rules of the language's morphology.

| Prefix | Suffix | Frequency | Examples |
|--------|--------|-----------|----------|
| `-ເ` | `-ນ` | 97 words | ເມືອງຕຸ້ມລານ, ເພື່ອຄັດຄ້ານ |
| `-ສ` | `-ນ` | 72 words | ສັດຈະກຸນ, ສ້າງອາຄານຮຽນ |
| `-ເ` | `-ງ` | 67 words | ເພາະຫຍັງ, ເຂດສັມພັນທະວົງ |
| `-ສ` | `-ດ` | 55 words | ສີດ, ສູນພັດທະນາຊົນນະບົດ |
| `-ແ` | `-ນ` | 52 words | ແກນ, ແຂວງສາລະວັນ |
| `-ອ` | `-ນ` | 48 words | ອຶນ, ອັສຈັນ |
| `-ເ` | `-ດ` | 47 words | ເມືອງຊົນນະບົດ, ເຟສບຸກແຟນເຜດ |
| `-ແ` | `-ງ` | 45 words | ແບັງ, ແມ່ນນັກການເມືອງ |
| `-ກ` | `-ນ` | 41 words | ກຽກກີດສະຖານ, ກິດຕິກອນຈະເລີນ |
| `-ເ` | `-າ` | 39 words | ເມືອງໄຊເຊດຖາ, ເຈົ້າເກດສະດາ |

### 6.5 Recursive Morpheme Segmentation

Using **Recursive Hierarchical Substitutability**, we decompose complex words into their constituent morphemes. This approach handles nested affixes (e.g., `prefix-prefix-root-suffix`).

| Word | Suggested Split | Confidence | Stem |
|------|-----------------|------------|------|
| ວັດສີພຸດທະບາດ | **`ວັດສີພຸດທະ-ບ-າດ`** | 7.5 | `ບ` |
| detectors | **`detector-s`** | 4.5 | `detector` |
| ການສືບສວນ | **`ການ-ສືບສວນ`** | 4.5 | `ສືບສວນ` |
| ການເຄື່ອນໄຫວ | **`ການ-ເຄື່ອນໄຫວ`** | 4.5 | `ເຄື່ອນໄຫວ` |
| olympians | **`olympian-s`** | 4.5 | `olympian` |
| recommended | **`recommend-ed`** | 4.5 | `recommend` |
| ministers | **`minister-s`** | 4.5 | `minister` |
| ການລົງທຶນ | **`ການ-ລົງທຶນ`** | 4.5 | `ລົງທຶນ` |
| ການເດີນທາງ | **`ການ-ເດີນທາງ`** | 4.5 | `ເດີນທາງ` |
| ການວິເຄາະ | **`ການ-ວິເຄາະ`** | 4.5 | `ວິເຄາະ` |
| ການໂຄສະນາ | **`ການ-ໂຄສະນາ`** | 4.5 | `ໂຄສະນາ` |
| republican | **`republic-an`** | 4.5 | `republic` |
| ໃນກາງຊຸມປີ | **`ໃນ-ກາງຊຸມປີ`** | 4.5 | `ກາງຊຸມປີ` |
| ອະທິນນາທານາ | **`ອະທິນນາທ-ານ-າ`** | 3.0 | `ອະທິນນາທ` |
| ການໄດ້ຍິນ | **`ການ-ໄດ້ຍິ-ນ`** | 3.0 | `ໄດ້ຍິ` |

### 6.6 Linguistic Interpretation

> **Automated Insight:**
The language Lao shows high morphological productivity. The subword models are significantly more efficient than word models, suggesting a rich system of affixation or compounding.

> **Note on Idiomaticity:** The high Idiomaticity Gap suggests a large number of frequent multi-word expressions or formulaic sequences that are statistically distinct from their component parts.

---
## 7. Summary & Recommendations

![Performance Dashboard](visualizations/performance_dashboard.png)

### Production Recommendations

| Component | Recommended | Rationale |
|-----------|-------------|-----------|
| Tokenizer | **64k BPE** | Best compression (4.28x) |
| N-gram | **2-gram** | Lowest perplexity (2,184) |
| Markov | **Context-4** | Highest predictability (98.3%) |
| Embeddings | **100d** | Balanced semantic capture and isotropy |


---
## Appendix: Metrics Glossary & Interpretation Guide

This section provides definitions, intuitions, and guidance for interpreting the metrics used throughout this report.

### Tokenizer Metrics

**Compression Ratio**
> *Definition:* The ratio of characters to tokens (chars/token). Measures how efficiently the tokenizer represents text.
>
> *Intuition:* Higher compression means fewer tokens needed to represent the same text, reducing sequence lengths for downstream models. A 3x compression means ~3 characters per token on average.
>
> *What to seek:* Higher is generally better for efficiency, but extremely high compression may indicate overly aggressive merging that loses morphological information.

**Average Token Length (Fertility)**
> *Definition:* Mean number of characters per token produced by the tokenizer.
>
> *Intuition:* Reflects the granularity of tokenization. Longer tokens capture more context but may struggle with rare words; shorter tokens are more flexible but increase sequence length.
>
> *What to seek:* Balance between 2-5 characters for most languages. Arabic/morphologically-rich languages may benefit from slightly longer tokens.

**Unknown Token Rate (OOV Rate)**
> *Definition:* Percentage of tokens that map to the unknown/UNK token, indicating words the tokenizer cannot represent.
>
> *Intuition:* Lower OOV means better vocabulary coverage. High OOV indicates the tokenizer encounters many unseen character sequences.
>
> *What to seek:* Below 1% is excellent; below 5% is acceptable. BPE tokenizers typically achieve very low OOV due to subword fallback.

### N-gram Model Metrics

**Perplexity**
> *Definition:* Measures how "surprised" the model is by test data. Mathematically: 2^(cross-entropy). Lower values indicate better prediction.
>
> *Intuition:* If perplexity is 100, the model is as uncertain as if choosing uniformly among 100 options at each step. A perplexity of 10 means effectively choosing among 10 equally likely options.
>
> *What to seek:* Lower is better. Perplexity decreases with larger n-grams (more context). Values vary widely by language and corpus size.

**Entropy**
> *Definition:* Average information content (in bits) needed to encode the next token given the context. Related to perplexity: perplexity = 2^entropy.
>
> *Intuition:* High entropy means high uncertainty/randomness; low entropy means predictable patterns. Natural language typically has entropy between 1-4 bits per character.
>
> *What to seek:* Lower entropy indicates more predictable text patterns. Entropy should decrease as n-gram size increases.

**Coverage (Top-K)**
> *Definition:* Percentage of corpus occurrences explained by the top K most frequent n-grams.
>
> *Intuition:* High coverage with few patterns indicates repetitive/formulaic text; low coverage suggests diverse vocabulary usage.
>
> *What to seek:* Depends on use case. For language modeling, moderate coverage (40-60% with top-1000) is typical for natural text.

### Markov Chain Metrics

**Average Entropy**
> *Definition:* Mean entropy across all contexts, measuring average uncertainty in next-word prediction.
>
> *Intuition:* Lower entropy means the model is more confident about what comes next. Context-1 has high entropy (many possible next words); Context-4 has low entropy (few likely continuations).
>
> *What to seek:* Decreasing entropy with larger context sizes. Very low entropy (<0.1) indicates highly deterministic transitions.

**Branching Factor**
> *Definition:* Average number of unique next tokens observed for each context.
>
> *Intuition:* High branching = many possible continuations (flexible but uncertain); low branching = few options (predictable but potentially repetitive).
>
> *What to seek:* Branching factor should decrease with context size. Values near 1.0 indicate nearly deterministic chains.

**Predictability**
> *Definition:* Derived metric: (1 - normalized_entropy) × 100%. Indicates how deterministic the model's predictions are.
>
> *Intuition:* 100% predictability means the next word is always certain; 0% means completely random. Real text falls between these extremes.
>
> *What to seek:* Higher predictability for text generation quality, but too high (>98%) may produce repetitive output.

### Vocabulary & Zipf's Law Metrics

**Zipf's Coefficient**
> *Definition:* The slope of the log-log plot of word frequency vs. rank. Zipf's law predicts this should be approximately -1.
>
> *Intuition:* A coefficient near -1 indicates the corpus follows natural language patterns where a few words are very common and most words are rare.
>
> *What to seek:* Values between -0.8 and -1.2 indicate healthy natural language distribution. Deviations may suggest domain-specific or artificial text.

**R² (Coefficient of Determination)**
> *Definition:* Measures how well the linear fit explains the frequency-rank relationship. Ranges from 0 to 1.
>
> *Intuition:* R² near 1.0 means the data closely follows Zipf's law; lower values indicate deviation from expected word frequency patterns.
>
> *What to seek:* R² > 0.95 is excellent; > 0.99 indicates near-perfect Zipf adherence typical of large natural corpora.

**Vocabulary Coverage**
> *Definition:* Cumulative percentage of corpus tokens accounted for by the top N words.
>
> *Intuition:* Shows how concentrated word usage is. If top-100 words cover 50% of text, the corpus relies heavily on common words.
>
> *What to seek:* Top-100 covering 30-50% is typical. Higher coverage indicates more repetitive text; lower suggests richer vocabulary.

### Word Embedding Metrics

**Isotropy**
> *Definition:* Measures how uniformly distributed vectors are in the embedding space. Computed as the ratio of minimum to maximum singular values.
>
> *Intuition:* High isotropy (near 1.0) means vectors spread evenly in all directions; low isotropy means vectors cluster in certain directions, reducing expressiveness.
>
> *What to seek:* Higher isotropy generally indicates better-quality embeddings. Values > 0.1 are reasonable; > 0.3 is good. Lower-dimensional embeddings tend to have higher isotropy.

**Average Norm**
> *Definition:* Mean magnitude (L2 norm) of word vectors in the embedding space.
>
> *Intuition:* Indicates the typical "length" of vectors. Consistent norms suggest stable training; high variance may indicate some words are undertrained.
>
> *What to seek:* Relatively consistent norms across models. The absolute value matters less than consistency (low std deviation).

**Cosine Similarity**
> *Definition:* Measures angular similarity between vectors, ranging from -1 (opposite) to 1 (identical direction).
>
> *Intuition:* Words with similar meanings should have high cosine similarity. This is the standard metric for semantic relatedness in embeddings.
>
> *What to seek:* Semantically related words should score > 0.5; unrelated words should be near 0. Synonyms often score > 0.7.

**t-SNE Visualization**
> *Definition:* t-Distributed Stochastic Neighbor Embedding - a dimensionality reduction technique that preserves local structure for visualization.
>
> *Intuition:* Clusters in t-SNE plots indicate groups of semantically related words. Spread indicates vocabulary diversity; tight clusters suggest semantic coherence.
>
> *What to seek:* Meaningful clusters (e.g., numbers together, verbs together). Avoid over-interpreting distances - t-SNE preserves local, not global, structure.

### General Interpretation Guidelines

1. **Compare within model families:** Metrics are most meaningful when comparing models of the same type (e.g., 8k vs 64k tokenizer).
2. **Consider trade-offs:** Better performance on one metric often comes at the cost of another (e.g., compression vs. OOV rate).
3. **Context matters:** Optimal values depend on downstream tasks. Text generation may prioritize different metrics than classification.
4. **Corpus influence:** All metrics are influenced by corpus characteristics. Wikipedia text differs from social media or literature.
5. **Language-specific patterns:** Morphologically rich languages (like Arabic) may show different optimal ranges than analytic languages.


### Visualizations Index

| Visualization | Description |
|---------------|-------------|
| Tokenizer Compression | Compression ratios by vocabulary size |
| Tokenizer Fertility | Average token length by vocabulary |
| Tokenizer OOV | Unknown token rates |
| Tokenizer Total Tokens | Total tokens by vocabulary |
| N-gram Perplexity | Perplexity by n-gram size |
| N-gram Entropy | Entropy by n-gram size |
| N-gram Coverage | Top pattern coverage |
| N-gram Unique | Unique n-gram counts |
| Markov Entropy | Entropy by context size |
| Markov Branching | Branching factor by context |
| Markov Contexts | Unique context counts |
| Zipf's Law | Frequency-rank distribution with fit |
| Vocab Frequency | Word frequency distribution |
| Top 20 Words | Most frequent words |
| Vocab Coverage | Cumulative coverage curve |
| Embedding Isotropy | Vector space uniformity |
| Embedding Norms | Vector magnitude distribution |
| Embedding Similarity | Word similarity heatmap |
| Nearest Neighbors | Similar words for key terms |
| t-SNE Words | 2D word embedding visualization |
| t-SNE Sentences | 2D sentence embedding visualization |
| Position Encoding | Encoding method comparison |
| Model Sizes | Storage requirements |
| Performance Dashboard | Comprehensive performance overview |

---
## About This Project

### Data Source

Models trained on [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly) - a monthly snapshot of Wikipedia articles across 300+ languages.

### Project

A project by **[Wikilangs](https://wikilangs.org)** - Open-source NLP models for every Wikipedia language.

### Maintainer

[Omar Kamali](https://omarkamali.com) - [Omneity Labs](https://omneitylabs.com)

### Citation

If you use these models in your research, please cite:

```bibtex
@misc{wikilangs2025,
  author = {Kamali, Omar},
  title = {Wikilangs: Open NLP Models for Wikipedia Languages},
  year = {2025},
  doi = {10.5281/zenodo.18073153},
  publisher = {Zenodo},
  url = {https://huggingface.co/wikilangs}
  institution = {Omneity Labs}
}
```

### License

MIT License - Free for academic and commercial use.

### Links

- 🌐 Website: [wikilangs.org](https://wikilangs.org)
- 🤗 Models: [huggingface.co/wikilangs](https://huggingface.co/wikilangs)
- 📊 Data: [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
- 👤 Author: [Omar Kamali](https://huggingface.co/omarkamali)
- 🤝 Sponsor: [Featherless AI](https://featherless.ai)
---
*Generated by Wikilangs Models Pipeline*

*Report Date: 2026-01-10 11:21:00*