Upload all models and assets for cdo (20251001)

README.md

@@ -23,14 +23,14 @@ dataset_info:
 metrics:
   - name: best_compression_ratio
     type: compression
-    value: 2.796
+    value: 2.892
   - name: best_isotropy
     type: isotropy
-    value: 0.5460
+    value: 0.5551
   - name: vocabulary_size
     type: vocab
-    value: 12714
-generated: 2025-12-28
+    value: 0
+generated: 2026-01-03
 ---
 
 # CDO - Wikilangs Models
@@ -44,12 +44,13 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 ### Models & Assets
 
 - Tokenizers (8k, 16k, 32k, 64k)
-- N-gram models (2, 3, 4-gram)
-- Markov chains (context of 1, 2, 3 and 4)
+- 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
+- Embeddings in various sizes and dimensions (aligned and unaligned)
 - Language Vocabulary
 - Language Statistics
+
 ![Performance Dashboard](visualizations/performance_dashboard.png)
 
 ### Analysis and Evaluation
@@ -59,7 +60,8 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 - [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
 - [4. Vocabulary Analysis](#4-vocabulary-analysis)
 - [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
-- [6. Summary & Recommendations](#6-summary--recommendations)
+- [6. Morphological Analysis (Experimental)](#6-morphological-analysis)
+- [7. Summary & Recommendations](#7-summary--recommendations)
 - [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
 - [Visualizations Index](#visualizations-index)
 
@@ -68,50 +70,49 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 
 ![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 |
 |------------|-------------|---------------|----------|--------------|
-| **32k** | 2.562x | 2.54 | 0.0007% | 298,320 |
-| **64k** | 2.796x 🏆 | 2.77 | 0.0007% | 273,367 |
+| **32k** | 2.757x | 2.76 | 0.1034% | 257,325 |
+| **64k** | 2.892x 🏆 | 2.90 | 0.1084% | 245,327 |
 
 ### Tokenization Examples
 
 Below are sample sentences tokenized with each vocabulary size:
 
-**Sample 1:** `Pender Gông (Ĭng-ngṳ̄: Pender County) sê Mī-guók North Carolina gì siŏh ciáh gôn...`
+**Sample 1:** `Bashkortostan sê Ngò̤-lò̤-sṳ̆ gì siŏh ciáh gê̤ṳng-huò-guók. gì gê̤ṳng-huò-guók`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 32k | `▁pen der ▁gông ▁( ĭng - ngṳ̄ : ▁pen der ... (+19 more)` | 29 |
-| 64k | `▁pender ▁gông ▁( ĭng - ngṳ̄ : ▁pender ▁county ) ... (+17 more)` | 27 |
-
-**Sample 2:** `Duâi dâi
+| 32k | `▁b ash k ort os tan ▁sê ▁ngò ̤- lò ... (+17 more)` | 27 |
+| 64k | `▁bash k ortos tan ▁sê ▁ngò ̤- lò ̤- sṳ̆ ... (+15 more)` | 25 |
 
-Chók-sié
-
-Guó-sié
-
-
-分類:1170 nièng-dâi`
+**Sample 2:** `Montague Gông (Ĭng-ngṳ̄: Montague County) sê Mī-guók Texas gì siŏh ciáh gông. gì...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 32k | `▁duâi ▁dâi ▁chók - sié ▁guó - sié ▁分類 : ... (+7 more)` | 17 |
-| 64k | `▁duâi ▁dâi ▁chók - sié ▁guó - sié ▁分類 : ... (+7 more)` | 17 |
+| 32k | `▁mont a gue ▁gông ▁( ĭng - ngṳ̄ : ▁mont ... (+16 more)` | 26 |
+| 64k | `▁montague ▁gông ▁( ĭng - ngṳ̄ : ▁montague ▁county ) ... (+12 more)` | 22 |
 
-**Sample 3:** `1000 nièng-dâi téng 1000 nièng 1 nguŏk 1 hô̤ kăi-sṳ̄, gáu 1009 nièng 12 nguŏk 31...`
+**Sample 3:** `Ochiltree Gông (Ĭng-ngṳ̄: Ochiltree County) sê Mī-guók Texas gì siŏh ciáh gông. ...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 32k | `▁ 1 0 0 0 ▁nièng - dâi ▁téng ▁ ... (+36 more)` | 46 |
-| 64k | `▁ 1 0 0 0 ▁nièng - dâi ▁téng ▁ ... (+36 more)` | 46 |
+| 32k | `▁o chi l t re e ▁gông ▁( ĭng - ... (+22 more)` | 32 |
+| 64k | `▁ochiltree ▁gông ▁( ĭng - ngṳ̄ : ▁ochiltree ▁county ) ... (+12 more)` | 22 |
 
 
 ### Key Findings
 
-- **Best Compression:** 64k achieves 2.796x compression
-- **Lowest UNK Rate:** 32k with 0.0007% unknown tokens
+- **Best Compression:** 64k achieves 2.892x compression
+- **Lowest UNK Rate:** 32k with 0.1034% unknown tokens
 - **Trade-off:** Larger vocabularies improve compression but increase model size
 - **Recommendation:** 32k vocabulary provides optimal balance for production use
 
@@ -120,57 +121,89 @@ Guó-sié
 
 ![N-gram Perplexity](visualizations/ngram_perplexity.png)
 
+![N-gram Unique](visualizations/ngram_unique.png)
+
 ![N-gram Coverage](visualizations/ngram_coverage.png)
 
 ### Results
 
-| N-gram | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
-|--------|------------|---------|----------------|------------------|-------------------|
-| **2-gram** | 2,092 🏆 | 11.03 | 13,738 | 34.2% | 70.6% |
-| **2-gram** | 517 🏆 | 9.01 | 13,773 | 57.4% | 92.0% |
-| **3-gram** | 6,902 | 12.75 | 35,914 | 23.0% | 49.3% |
-| **3-gram** | 2,154 | 11.07 | 33,837 | 33.3% | 72.5% |
-| **4-gram** | 16,500 | 14.01 | 75,913 | 16.0% | 37.8% |
-| **4-gram** | 6,830 | 12.74 | 94,271 | 22.4% | 53.8% |
+| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
+|--------|---------|------------|---------|----------------|------------------|-------------------|
+| **2-gram** | Word | 3,105 | 11.60 | 11,679 | 27.7% | 59.2% |
+| **2-gram** | Subword | 342 🏆 | 8.42 | 6,912 | 63.5% | 95.8% |
+| **3-gram** | Word | 4,698 | 12.20 | 17,954 | 23.8% | 52.2% |
+| **3-gram** | Subword | 1,659 | 10.70 | 21,000 | 36.1% | 75.8% |
+| **4-gram** | Word | 8,483 | 13.05 | 30,938 | 18.6% | 45.4% |
+| **4-gram** | Subword | 5,744 | 12.49 | 69,193 | 23.7% | 55.8% |
 
 ### Top 5 N-grams by Size
 
-**2-grams:**
+**2-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `gì siŏh` | 6,258 |
+| 2 | `siŏh ciáh` | 6,232 |
+| 3 | `mī guók` | 3,385 |
+| 4 | `sê mī` | 3,191 |
+| 5 | `gì gông` | 3,000 |
+
+**3-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `gì siŏh ciáh` | 5,413 |
+| 2 | `sê mī guók` | 3,173 |
+| 3 | `siŏh ciáh gông` | 3,000 |
+| 4 | `ciáh gông gì` | 2,557 |
+| 5 | `gông gì gông` | 2,557 |
+
+**4-grams (Word):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `分類 :` | 17,792 |
-| 2 | `̤ ng` | 9,653 |
-| 3 | `. 分類` | 8,000 |
-| 4 | `- guók` | 7,750 |
-| 5 | `- sié` | 7,747 |
+| 1 | `gì siŏh ciáh gông` | 3,000 |
+| 2 | `ciáh gông gì gông` | 2,557 |
+| 3 | `siŏh ciáh gông gì` | 2,557 |
+| 4 | `county sê mī guók` | 1,971 |
+| 5 | `gông sê mī guók` | 1,029 |
 
-**3-grams:**
+**2-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `. 分類 :` | 8,000 |
-| 2 | `gì siŏh ciáh` | 5,565 |
-| 3 | `- ngṳ ̄` | 4,336 |
-| 4 | `mī - guók` | 3,641 |
-| 5 | `gâe ̤ ng` | 3,480 |
+| 1 | `n g` | 146,797 |
+| 2 | `_ g` | 59,970 |
+| 3 | `g -` | 55,946 |
+| 4 | `g _` | 55,139 |
+| 5 | `_ s` | 41,311 |
 
-**4-grams:**
+**3-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `sê mī - guók` | 3,211 |
-| 2 | `gì siŏh ciáh gông` | 3,000 |
-| 3 | `ciáh gông . 分類` | 3,000 |
-| 4 | `gông . 分類 :` | 3,000 |
-| 5 | `siŏh ciáh gông .` | 3,000 |
+| 1 | `n g -` | 55,920 |
+| 2 | `n g _` | 55,025 |
+| 3 | `_ g ì` | 23,090 |
+| 4 | `g ì _` | 22,312 |
+| 5 | `_ s i` | 14,134 |
+
+**4-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `_ g ì _` | 22,161 |
+| 2 | `_ s ê _` | 13,231 |
+| 3 | `n g _ g` | 11,336 |
+| 4 | `i ŏ h _` | 10,632 |
+| 5 | `_ s i ŏ` | 9,391 |
 
 
 ### Key Findings
 
-- **Best Perplexity:** 2-gram with 517
+- **Best Perplexity:** 2-gram (subword) with 342
 - **Entropy Trend:** Decreases with larger n-grams (more predictable)
-- **Coverage:** Top-1000 patterns cover ~54% of corpus
+- **Coverage:** Top-1000 patterns cover ~56% of corpus
 - **Recommendation:** 4-gram or 5-gram for best predictive performance
 
 ---
@@ -178,55 +211,86 @@ Guó-sié
 
 ![Markov Entropy](visualizations/markov_entropy.png)
 
+![Markov Contexts](visualizations/markov_contexts.png)
+
 ![Markov Branching](visualizations/markov_branching.png)
 
 ### Results
 
-| Context | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
-|---------|-------------|------------|------------------|-----------------|----------------|
-| **1** | 0.2803 | 1.214 | 3.49 | 48,699 | 72.0% |
-| **1** | 0.3942 | 1.314 | 4.02 | 31,614 | 60.6% |
-| **2** | 0.1991 | 1.148 | 1.83 | 169,503 | 80.1% |
-| **2** | 0.3616 | 1.285 | 2.00 | 127,156 | 63.8% |
-| **3** | 0.1556 | 1.114 | 1.42 | 308,939 | 84.4% |
-| **3** | 0.2179 | 1.163 | 1.54 | 253,902 | 78.2% |
-| **4** | 0.0983 🏆 | 1.071 | 1.21 | 437,205 | 90.2% |
-| **4** | 0.1764 🏆 | 1.130 | 1.38 | 389,634 | 82.4% |
+| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
+|---------|---------|-------------|------------|------------------|-----------------|----------------|
+| **1** | Word | 0.4882 | 1.403 | 4.73 | 29,670 | 51.2% |
+| **1** | Subword | 0.3461 | 1.271 | 2.92 | 25,622 | 65.4% |
+| **2** | Word | 0.3187 | 1.247 | 1.80 | 139,308 | 68.1% |
+| **2** | Subword | 0.2753 | 1.210 | 1.79 | 74,780 | 72.5% |
+| **3** | Word | 0.1201 | 1.087 | 1.23 | 249,012 | 88.0% |
+| **3** | Subword | 0.2343 | 1.176 | 1.69 | 133,665 | 76.6% |
+| **4** | Word | 0.0526 🏆 | 1.037 | 1.09 | 301,670 | 94.7% |
+| **4** | Subword | 0.2290 | 1.172 | 1.53 | 225,577 | 77.1% |
 
-### Generated Text Samples
+### Generated Text Samples (Word-based)
 
-Below are text samples generated from each Markov chain model:
+Below are text samples generated from each word-based Markov chain model:
 
 **Context Size 1:**
 
-1. `- hū siék gì siŏh ciáh gông . 分類 : chĭng - uăng - pū -`
-2. `̤ k nâ sáng ĕu - ngiòng ( 螺洲路 ) guōng - dŏng - dōi -`
-3. `gì dâ ̤ 18 艭 ngiê - guók - guó - dók “ . chók -`
+1. `gì hiŏng 蘆洋鄉 bìng dēng có̤i kăi sṳ̄ gó nâ háng nè̤ng gó̤ lō̤ 法老 gái`
+2. `sê dṳ̆ng huà ìng chê 邢臺市 lòng cĭ 退之 sê ĕng diŏh adelaide ô 2 nguŏk`
+3. `siŏh bĭk cék éng sáuk īng lĭk â̤ dé̤ṳng buŏng nàng áng cê sê turkic ngṳ̄`
 
 **Context Size 2:**
 
-1. `分類 : 1370 nièng - dâi gì lùng - dŭng - ngŏk liù - giù - dôi`
-2. `̤ ng hók - gióng , dâi - biēu gê ̤ ṳng - sāng - dōng gâe`
-3. `. 分類 : 200 nièng - dâi - mā 分類 : 1300年代`
+1. `gì siŏh cṳ̄ng lòi ĭng ôi sĭng ô 79 ciáh mŭk sṳ̆ 佳木斯 sê dṳ̆ng guók sĭng`
+2. `siŏh ciáh chê hăk kṳ̆ 市轄區 lĭk sṳ̄ diē sié sê siŏh cṳ̄ng ciŏng muòng dò̤ lā̤`
+3. `mī guók montana gì siŏh ciáh gông gì gông`
 
 **Context Size 3:**
 
-1. `. 分類 : minnesota gì gông`
-2. `gì siŏh ciáh dê - ngék - chê . 分類 : hù - báe ̤ k - chiă`
-3. `- ngṳ ̄ : lafayette county ) sê mī - guók gì buô - hông gì sṳ ̆`
+1. `gì siŏh ciáh duâi kṳ̆ gì duâi kṳ̆`
+2. `sê mī guók kansas gì siŏh ciáh cê mō̤ diŏh eta gì piăng âu gâe̤ng sigma gì sèng`
+3. `siŏh ciáh gông gì gông`
 
 **Context Size 4:**
 
-1. `sê mī - guók colorado gì siŏh ciáh gông . 分類 : florida gì gông`
-2. `gông . 分類 : michigan gì gông`
-3. `siŏh ciáh gông . 分類 : indiana gì gông`
+1. `gì siŏh ciáh gông gì gông`
+2. `siŏh ciáh gông gì gông`
+3. `county sê mī guók nebraska gì siŏh ciáh gông gì gông`
+
+
+### Generated Text Samples (Subword-based)
+
+Below are text samples generated from each subword-based Markov chain model:
+
+**Context Size 1:**
+
+1. `_ônià_hô̤_sênty)_`
+2. `gì_ciônièng,_sṳ̄:`
+3. `ngăngôner_g-ccáh`
+
+**Context Size 2:**
+
+1. `nguô_ka_cĭ_gì_dâ̤_`
+2. `_gì_sié-sê_hŏk-cê`
+3. `g-hĭ_(獨聯體)_gông_n`
+
+**Context Size 3:**
+
+1. `ng-hèng_biéng,_mac`
+2. `ng_adahoma_gì_(兩個聲`
+3. `_gì_siōng-dĕ̤ng-ŭk_`
+
+**Context Size 4:**
+
+1. `_gì_«sṳ̀ng-kṳ̆_dĕk-bi`
+2. `_sê_„發現更大的世界“)_có̤_c`
+3. `ng_găk_chăng-muò_(𧋘`
 
 
 ### Key Findings
 
-- **Best Predictability:** Context-4 with 90.2% predictability
+- **Best Predictability:** Context-4 (word) with 94.7% predictability
 - **Branching Factor:** Decreases with context size (more deterministic)
-- **Memory Trade-off:** Larger contexts require more storage (389,634 contexts)
+- **Memory Trade-off:** Larger contexts require more storage (225,577 contexts)
 - **Recommendation:** Context-3 or Context-4 for text generation
 
 ---
@@ -242,64 +306,64 @@ Below are text samples generated from each Markov chain model:
 
 | Metric | Value |
 |--------|-------|
-| Vocabulary Size | 12,714 |
-| Total Tokens | 590,881 |
-| Mean Frequency | 46.47 |
+| Vocabulary Size | 9,559 |
+| Total Tokens | 467,385 |
+| Mean Frequency | 48.89 |
 | Median Frequency | 3 |
-| Frequency Std Dev | 447.20 |
+| Frequency Std Dev | 395.71 |
 
 ### Most Common Words
 
 | Rank | Word | Frequency |
 |------|------|-----------|
-| 1 | gì | 24,268 |
-| 2 | 分類 | 17,794 |
-| 3 | ng | 16,472 |
-| 4 | sê | 15,967 |
-| 5 | siŏh | 9,713 |
-| 6 | guók | 9,302 |
-| 7 | gông | 9,087 |
-| 8 | sié | 8,595 |
-| 9 | nièng | 7,825 |
-| 10 | dâi | 7,699 |
+| 1 | gì | 23,295 |
+| 2 | sê | 14,068 |
+| 3 | siŏh | 9,247 |
+| 4 | gông | 9,087 |
+| 5 | guók | 8,549 |
+| 6 | ciáh | 7,131 |
+| 7 | nièng | 5,854 |
+| 8 | ngṳ̄ | 5,277 |
+| 9 | sié | 4,616 |
+| 10 | gáu | 4,179 |
 
 ### Least Common Words (from vocabulary)
 
 | Rank | Word | Frequency |
 |------|------|-----------|
-| 1 | 燈泡厰 | 2 |
-| 2 | 搪瓷厰 | 2 |
-| 3 | 保溫瓶厰 | 2 |
-| 4 | 啤酒厰 | 2 |
-| 5 | 福大機械厰 | 2 |
-| 6 | 抗生素厰 | 2 |
-| 7 | kbo | 2 |
-| 8 | 우주항공청 | 2 |
-| 9 | cho | 2 |
-| 10 | chit | 2 |
+| 1 | woolridge | 2 |
+| 2 | imperiyası | 2 |
+| 3 | abş | 2 |
+| 4 | çox | 2 |
+| 5 | dünyada | 2 |
+| 6 | bütün | 2 |
+| 7 | 嘉祿 | 2 |
+| 8 | 六一路 | 2 |
+| 9 | 神壇樹 | 2 |
+| 10 | 신단수 | 2 |
 
 ### Zipf's Law Analysis
 
 | Metric | Value |
 |--------|-------|
 | Zipf Coefficient | 1.3995 |
-| R² (Goodness of Fit) | 0.979429 |
+| R² (Goodness of Fit) | 0.957431 |
 | Adherence Quality | **excellent** |
 
 ### Coverage Analysis
 
 | Top N Words | Coverage |
 |-------------|----------|
-| Top 100 | 55.6% |
-| Top 1,000 | 90.8% |
-| Top 5,000 | 97.1% |
-| Top 10,000 | 99.1% |
+| Top 100 | 52.2% |
+| Top 1,000 | 91.7% |
+| Top 5,000 | 98.0% |
+| Top 10,000 | 0.0% |
 
 ### Key Findings
 
-- **Zipf Compliance:** R²=0.9794 indicates excellent adherence to Zipf's law
-- **High Frequency Dominance:** Top 100 words cover 55.6% of corpus
-- **Long Tail:** 2,714 words needed for remaining 0.9% coverage
+- **Zipf Compliance:** R²=0.9574 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 52.2% of corpus
+- **Long Tail:** -441 words needed for remaining 100.0% coverage
 
 ---
 ## 5. Word Embeddings Evaluation
@@ -312,24 +376,80 @@ Below are text samples generated from each Markov chain model:
 
 ![t-SNE Sentences](visualizations/tsne_sentences.png)
 
-### Model Comparison
 
-| Model | Vocab Size | Dimension | Avg Norm | Std Norm | Isotropy |
-|-------|------------|-----------|----------|----------|----------|
-| **mono_32d** | 7,009 | 32 | 4.149 | 1.118 | 0.5460 🏆 |
-| **mono_64d** | 7,009 | 64 | 4.243 | 1.106 | 0.2037 |
-| **mono_128d** | 7,009 | 128 | 4.233 | 1.119 | 0.0381 |
-| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+### 5.1 Cross-Lingual Alignment
+
+> *Note: Multilingual alignment visualization not available for this language.*
+
+
+### 5.2 Model Comparison
+
+| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
+|-------|-----------|----------|------------------|---------------|----------------|
+| **mono_32d** | 32 | 0.5551 🏆 | 0.4156 | N/A | N/A |
+| **mono_64d** | 64 | 0.1856 | 0.4055 | N/A | N/A |
+| **mono_128d** | 128 | 0.0279 | 0.4128 | N/A | N/A |
 
 ### Key Findings
 
-- **Best Isotropy:** mono_32d with 0.5460 (more uniform distribution)
-- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
-- **Vocabulary Coverage:** All models cover 7,009 words
-- **Recommendation:** 100d for balanced semantic capture and efficiency
+- **Best Isotropy:** mono_32d with 0.5551 (more uniform distribution)
+- **Semantic Density:** Average pairwise similarity of 0.4113. Lower values indicate better semantic separation.
+- **Alignment Quality:** No aligned models evaluated in this run.
+- **Recommendation:** 128d aligned for best cross-lingual performance
+
+---
+## 6.  Morphological Analysis (Experimental)
+
+> ⚠️ **Warning:** This language shows low morphological productivity. The statistical signals used for this analysis may be noisy or less reliable than for morphologically rich languages.
+
+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 | **0.000** | Low morphological productivity | ⚠️ Likely unreliable |
+| Idiomaticity Gap | **-1.000** | Low formulaic 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.
+
+*No productive affixes detected.*
+
+
+### 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 |
+|------|----------|------------------|----------|
+| `áung` | 1.97x | 9 contexts | táung, láung, dáung |
+| `âung` | 1.96x | 9 contexts | câung, bâung, hâung |
+| `iăng` | 1.80x | 7 contexts | hiăng, siăng, giăng |
+| `iāng` | 1.55x | 8 contexts | liāng, biāng, ciāng |
+
+### 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.
+
+*No significant affix co-occurrences detected.*
+
+
+### 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`).
+
+*Insufficient data for recursive segmentation.*
+
+
+### 6.6 Linguistic Interpretation
+
+> **Automated Insight:**
+The language CDO appears to be more isolating or has a highly fixed vocabulary. Word-level models perform nearly as well as subword models, indicating fewer productive morphological processes.
 
 ---
-## 6. Summary & Recommendations
+## 7. Summary & Recommendations
 
 ![Performance Dashboard](visualizations/performance_dashboard.png)
 
@@ -337,11 +457,12 @@ Below are text samples generated from each Markov chain model:
 
 | Component | Recommended | Rationale |
 |-----------|-------------|-----------|
-| Tokenizer | **32k BPE** | Best compression (2.80x) with low UNK rate |
-| N-gram | **5-gram** | Lowest perplexity (517) |
-| Markov | **Context-4** | Highest predictability (90.2%) |
+| Tokenizer | **64k BPE** | Best compression (2.89x) |
+| N-gram | **2-gram** | Lowest perplexity (342) |
+| Markov | **Context-4** | Highest predictability (94.7%) |
 | Embeddings | **100d** | Balanced semantic capture and isotropy |
 
+
 ---
 ## Appendix: Metrics Glossary & Interpretation Guide
 
@@ -531,7 +652,8 @@ If you use these models in your research, please cite:
   author = {Kamali, Omar},
   title = {Wikilangs: Open NLP Models for Wikipedia Languages},
   year = {2025},
-  publisher = {HuggingFace},
+  doi = {10.5281/zenodo.18073153},
+  publisher = {Zenodo},
   url = {https://huggingface.co/wikilangs}
   institution = {Omneity Labs}
 }
@@ -547,7 +669,8 @@ MIT License - Free for academic and commercial use.
 - 🤗 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: 2025-12-28 16:25:16*
+*Report Date: 2026-01-03 09:43:04*

models/embeddings/monolingual/cdo_128d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87273b834175c8ffca7ff96a3a21f5890c33eb81fd8bd600e3f7749a3c1efde6
-size 1031314655
+oid sha256:8c366500bae149b287db9cc735aecf709e56990e0f1dd949aa01f743c5b542bf
+size 1030106435

models/embeddings/monolingual/cdo_128d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 128,
   "version": "monolingual",
   "training_params": {
-    "dim": 128,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 128
   },
-  "vocab_size": 7009
+  "vocab_size": 5854
 }

models/embeddings/monolingual/cdo_32d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d65b7a0893696382e676534c0654d0fa33b57a4337091da15492cd7c0abf3b48
-size 257931743
+oid sha256:9ee34c3106d7321309452b983807ebe84c3a153af734404b31b03ac0f6076d5c
+size 257610563

models/embeddings/monolingual/cdo_32d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 32,
   "version": "monolingual",
   "training_params": {
-    "dim": 32,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 32
   },
-  "vocab_size": 7009
+  "vocab_size": 5854
 }

models/embeddings/monolingual/cdo_64d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70424ab1f058bf1eec92ffa8c17229b3451d02319f7295d7a82908645320bdb6
-size 515726047
+oid sha256:aad5fc0dca0e268941025dbedc1373061213d3a2602f82f5fb36a6a8ac4c5600
+size 515109187

models/embeddings/monolingual/cdo_64d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 64,
   "version": "monolingual",
   "training_params": {
-    "dim": 64,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 64
   },
-  "vocab_size": 7009
+  "vocab_size": 5854
 }

models/subword_markov/cdo_markov_ctx1_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0f235b61b31e910d9f0a63bff6431ca1b061d7d60c93f009c3efb2d4dd4d8d94
-size 859391
+oid sha256:41c23bda939ee89e9ea22c7d220f77ade97cf29b5706a8ccea6a5a1649c35d5c
+size 588804

models/subword_markov/cdo_markov_ctx1_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "subword",
   "language": "cdo",
-  "unique_contexts": 31614,
-  "total_transitions": 2905389
+  "unique_contexts": 25622,
+  "total_transitions": 2209713
 }

models/subword_markov/cdo_markov_ctx2_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95c5781711001d90b8d98035bad77dc2eafa9ca4d93f92cbe4e187b3ca6f467c
-size 2273669
+oid sha256:635b5359d40f1fb9f72a9835efb0cf9b06c7ece40db0ae41691aa56bcf42a68f
+size 1373275

models/subword_markov/cdo_markov_ctx2_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "subword",
   "language": "cdo",
-  "unique_contexts": 127156,
-  "total_transitions": 2888681
+  "unique_contexts": 74780,
+  "total_transitions": 2199286
 }

models/subword_markov/cdo_markov_ctx3_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0112fc9e03100831bb45b13b125d60693af2b6111630b553ae0b8802b325840d
-size 4278912
+oid sha256:edbc276446953ce254374098ddc144b562d74f64b6dc6b037712625c3474bcc7
+size 2410456

models/subword_markov/cdo_markov_ctx3_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "subword",
   "language": "cdo",
-  "unique_contexts": 253902,
-  "total_transitions": 2871973
+  "unique_contexts": 133665,
+  "total_transitions": 2188859
 }

models/subword_markov/cdo_markov_ctx4_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:863a8d123baa32acad56367516fba40e4df9fe4d98e2a22ce9e27a3dcadf37ff
-size 6583946
+oid sha256:73b83a95594a9de41178ee150977a694d8fe0fc9528952db1f4037278640135b
+size 3940722

models/subword_markov/cdo_markov_ctx4_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 4,
   "variant": "subword",
   "language": "cdo",
-  "unique_contexts": 389634,
-  "total_transitions": 2855265
+  "unique_contexts": 225577,
+  "total_transitions": 2178432
 }

models/subword_ngram/cdo_2gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:429757e59b20addd0ca423c66011cf245936cb430796406a7e56ac432fbc78ff
-size 180569
+oid sha256:07438dbbdf7df68795d2567c55c6704ba9e2aa37a1e974f3796653e4ee46d715
+size 93187

models/subword_ngram/cdo_2gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 2,
   "variant": "subword",
   "language": "cdo",
-  "unique_ngrams": 13773,
-  "total_ngrams": 2905389
+  "unique_ngrams": 6912,
+  "total_ngrams": 2209713
 }

models/subword_ngram/cdo_3gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb21e2e89d34f0d79e0ad314920dfc3371d0dbdc72b20e5448b52c64d94e3daf
-size 481494
+oid sha256:9bc12fdb56f63acfec879929e218e1b08b69ed6bc6caaf23de5405818299a15a
+size 290607

models/subword_ngram/cdo_3gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 3,
   "variant": "subword",
   "language": "cdo",
-  "unique_ngrams": 33837,
-  "total_ngrams": 2888681
+  "unique_ngrams": 21000,
+  "total_ngrams": 2199286
 }

models/subword_ngram/cdo_4gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f82d92e0227f12077424a14d88f6d50fbac2bc1d12b803abe7c34c6bc8a1e7e
-size 1234369
+oid sha256:d7bbf6e008db71c8cab5f01f7b94dece221cc12e30b6c0bbde60267e28ad3d1b
+size 893634

models/subword_ngram/cdo_4gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 4,
   "variant": "subword",
   "language": "cdo",
-  "unique_ngrams": 94271,
-  "total_ngrams": 2871973
+  "unique_ngrams": 69193,
+  "total_ngrams": 2188859
 }

models/tokenizer/cdo_tokenizer_32k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:29cc26ce42341f87bf600b104422987383bb16aab206d714d8025b627a81318a
-size 652498
+oid sha256:e4ea04b7b5ac9eb6c0c47f8743b9ec573da4a66b11329b62767916c896096283
+size 659402

models/tokenizer/cdo_tokenizer_64k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fd9f7c3b27cd6d59d2600b4da0253681a1c8987f94a736a1458e336f86ae632c
-size 1220234
+oid sha256:543fc6291e9f68ca726ae3c62da2bf8a9f2e8964f61ac8c8f47e7620b30680d4
+size 1252522

models/vocabulary/cdo_vocabulary.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f67094c35332c0e243a9e0fa3b227560d5e7cbeb597f1f5f91eeaae915d87402
-size 220074
+oid sha256:13d428e7f371fae4d1817f7fd0cad404bb08c4eebfeba6b19d2159ec0858966d
+size 161421

models/vocabulary/cdo_vocabulary_metadata.json

@@ -1,16 +1,17 @@
 {
   "language": "cdo",
-  "vocabulary_size": 12714,
+  "vocabulary_size": 9559,
+  "variant": "full",
   "statistics": {
-    "type_token_ratio": 0.07780709085921002,
+    "type_token_ratio": 0.06153581253100364,
     "coverage": {
-      "top_100": 0.5244804991801553,
-      "top_1000": 0.8557711325341176,
-      "top_5000": 0.9148462073409597,
-      "top_10000": 0.9338142876283201
+      "top_100": 0.4998278145152364,
+      "top_1000": 0.8788593121599849,
+      "top_5000": 0.9388721030817102,
+      "top_10000": 0.9589624594646672
     },
-    "hapax_count": 36067,
-    "hapax_ratio": 0.7393657366597651,
-    "total_documents": 16708
+    "hapax_count": 20461,
+    "hapax_ratio": 0.6815789473684211,
+    "total_documents": 10427
   }
 }

models/word_markov/cdo_markov_ctx1_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:216d791b7daa8b64656b1cba78bbbc02aa7f5ae7d9bf1ce3a3acf8fc6ca30ad8
-size 2218064
+oid sha256:b7c52cf701393a62ef55eb76e071bcc146cb575dcba8d7de8f629c6f59a0ffa9
+size 1378282

models/word_markov/cdo_markov_ctx1_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "word",
   "language": "cdo",
-  "unique_contexts": 48699,
-  "total_transitions": 989207
+  "unique_contexts": 29670,
+  "total_transitions": 477419
 }

models/word_markov/cdo_markov_ctx2_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59c0afaec27cafdf472078114f0c1e6330cbc442a0cf86a0de739eb654009647
-size 4081907
+oid sha256:4e2d34224b73bb7d088188c6354a0a9b2ed3b721bfecbe369bd09f27bc4bf8eb
+size 3117924

models/word_markov/cdo_markov_ctx2_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "word",
   "language": "cdo",
-  "unique_contexts": 169503,
-  "total_transitions": 972499
+  "unique_contexts": 139308,
+  "total_transitions": 466992
 }

models/word_markov/cdo_markov_ctx3_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9a9fad10b27aa485ee443b1b3fb11cf1c68941f562e62e4a5d392900045e177d
-size 6371893
+oid sha256:751ecd29db06623ef008eb11136a6de55e6177de5d5106326042cf155bf67096
+size 4876570

models/word_markov/cdo_markov_ctx3_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "word",
   "language": "cdo",
-  "unique_contexts": 308939,
-  "total_transitions": 955791
+  "unique_contexts": 249012,
+  "total_transitions": 456565
 }

models/word_markov/cdo_markov_ctx4_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e117de2ccb82cee04ee59547125507223f105436e0ac8437360575ade999d100
-size 8441712
+oid sha256:be6fc8ceb5ad611ae4227c2c3e9eb01458860d29280986707129fb49af2ed1fc
+size 5983737

models/word_markov/cdo_markov_ctx4_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 4,
   "variant": "word",
   "language": "cdo",
-  "unique_contexts": 437205,
-  "total_transitions": 939084
+  "unique_contexts": 301670,
+  "total_transitions": 446138
 }

models/word_ngram/cdo_2gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b2869b3ed62a803296d4b64cbd8fba21a1774ee3355a39d11e555520e9c5161b
-size 191655
+oid sha256:1665b5acdc0653c9ed337ac32fede9cd787b10dd4061bd94b629fa966ee9aaf1
+size 178246

models/word_ngram/cdo_2gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 2,
   "variant": "word",
   "language": "cdo",
-  "unique_ngrams": 13738,
-  "total_ngrams": 989207
+  "unique_ngrams": 11679,
+  "total_ngrams": 477419
 }

models/word_ngram/cdo_3gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5bfe056eb57d732812afef695d46bb07dbebf064ff917dfb6ec89e7a9cdec64d
-size 547998
+oid sha256:0dbe777e25174fb6b91ddeadf4d3f6969ac19328aeeb43492d068cc47fffedb9
+size 311342

models/word_ngram/cdo_3gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 3,
   "variant": "word",
   "language": "cdo",
-  "unique_ngrams": 35914,
-  "total_ngrams": 972499
+  "unique_ngrams": 17954,
+  "total_ngrams": 466992
 }

models/word_ngram/cdo_4gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c9d589fedb9dc7dce0fa306987fb5585c1f152679ee6c6bf37390de81fc147cc
-size 1141850
+oid sha256:11d7a87e924b2a61e7e8acb9a1eb51754eb4c044273e29e85c59b97269584a38
+size 556300

models/word_ngram/cdo_4gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 4,
   "variant": "word",
   "language": "cdo",
-  "unique_ngrams": 75913,
-  "total_ngrams": 955791
+  "unique_ngrams": 30938,
+  "total_ngrams": 456565
 }