README.md

---

license: cc-by-4.0
language:
- ba
pretty_name: Bashkir Frequency Index (152M tokens)
task_categories:
- text-classification
configs:
- config_name: public
  data_files:
  - split: train
    path: frequency_public.parquet
- config_name: full
  data_files:
  - split: train
    path: frequency_full.parquet
tags:
- bashkir
- frequency
- wordlist
- lexicon
- low-resource
- cyrillic
- corpus-statistics
---


# Bashkir Frequency Index v13.0-af1

Word-frequency index for the Bashkir language, computed over **152.16 million tokens** of clean monolingual text across diverse public domains (periodicals, news media, literary publications, encyclopedic texts, books, and general web archives). Non-Bashkir language admixture and scanning artifacts were filtered using automated language-filtering pipelines.

## Configurations

| Config | Rows | Cutoff | Use Case |
|---|---:|---:|---|
| **`public`** (recommended) | **589,980** | count ≥ 3 | NLP applications, spellchecking, text processing, linguistic research |
| **`full`** | **801,775** | count ≥ 2 | Comprehensive morphology studies, lexical research |

Additionally provided in the repository:
- [`top_coverage_50.json`](top_coverage_50.json): the minimal set of **1,177 word forms covering 50% of all tokens** in the corpus.
- [`top_500.json`](top_500.json): top-500 core frequency vocabulary list.

## Schema

| Column | Type | Description |
|---|---|---|
| `rank` | int64 | Position by descending frequency |
| `word` | string | Word form, lowercase Cyrillic |
| `count` | int64 | Absolute frequency count (occurrences) |
| `doc_freq` | int64 | Number of distinct sentences containing this word form |
| `freq_per_million` | double | Normalized instances per million tokens (FPM / IPM) |

## Sample (Top 10 Words)

| Rank | Word | Function / Translation | Count | FPM | Doc Frequency |
|---:|---|---|---:|---:|---:|
| 1 | **һәм** | conjunction "and" | 2,241,951 | 14,734.0 | 1,912,898 |
| 2 | **менән** | postposition "with" | 1,777,603 | 11,682.3 | 1,655,420 |
| 3 | **бер** | numeral "one / a certain" | 1,086,312 | 7,139.2 | 1,003,231 |
| 4 | **ла** | particle "also / and" | 973,753 | 6,399.5 | 911,626 |
| 5 | **ул** | pronoun "he / she / it / that" | 849,049 | 5,579.9 | 831,161 |
| 6 | **өсөн** | pronoun "this" | 830,427 | 5,457.5 | 781,355 |
| 7 | **тип** | postposition "for / sake of" | 828,274 | 5,443.4 | 799,514 |
| 8 | **был** | quotative gerund "saying / as" | 827,544 | 5,438.6 | 815,399 |
| 9 | **лә** | particle "also / and" | 602,419 | 3,959.1 | 569,728 |
| 10 | **буйынса** | postposition/link "according to" | 513,002 | 3,371.4 | 482,030 |

## Token Coverage

| Tokens Covered | Forms Needed |
|---:|---:|
| **25%** | **148** |
| **50%** | **1,177** |
| **80%** | **11,265** |
| **90%** | **32,544** |
| **95%** | **76,326** |

## Method

- **Tokenization:** Regex lowercase tokenization over Cyrillic script including all 9 Bashkir-specific letters (`Ә Ғ Ҙ Ҡ Ң Ө Ҫ Ү Һ`).
- **Corpus Base:** 12.89M sentences (152.16M tokens).
- **Filtering:** Automated function-word language filtering to remove foreign language leaks.
- **Form Constraints:** Isolated single-character forms removed (except `ә`); concatenated strings (>24 chars) removed.

## Loading

```python

from datasets import load_dataset



# Load recommended public configuration

dataset = load_dataset("failed09/bashkir-frequency-index", "public")

df = dataset["train"].to_pandas()

print(df.head(10))

```

## Citation

```bibtex

@dataset{failed09_bashkir_frequency_index_2026,

  title     = {Bashkir Frequency Index v13.0-af1},

  author    = {failed09},

  year      = {2026},

  publisher = {Hugging Face},

  url       = {https://huggingface.co/datasets/failed09/bashkir-frequency-index}

}

```

## Open Bashkir Data and Sources 🐝

This dataset is part of an open-source effort to support the development,
preservation and practical use of the Bashkir language. Other related models,
datasets and tools are available on the author's Hugging Face profile.

The author does not claim ownership or authorship of the source texts or other
training materials used to derive this index. Rights and licensing remain with
the original authors, publishers and dataset providers. Source texts are not
redistributed in this repository; users should review and follow the licenses
and attribution requirements of the relevant upstream resources.