# CHiPS: Character Histograms and Positional Signals for Lightweight Authorship Attribution in Romanian Texts

arXiv: https://arxiv.org/abs/2607.22884 · HTML: https://mlsanity.com/paper/2607.22884/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 3 of its 28 references that are in the map.
- First posted on arXiv 24 Jul 2026.

## Builds on (most cited first, 3 of 3)

- [A comparison of several AI techniques for authorship attribution on Romanian texts](https://mlsanity.com/paper/2211.05180/) (2022)
- [BERT-based Authorship Attribution on the Romanian Dataset called ROST](https://mlsanity.com/paper/2301.12500/) (2023)
- [Oldies but Goldies: The Potential of Character N-grams for Romanian Texts](https://mlsanity.com/paper/2506.15650/) (2025)

## Abstract (arXiv)

We propose CHiPS, a lightweight character-level authorship attribution method for Romanian texts. All reported experiments are closed-set: the true author is one of the candidate authors in the training data. CHiPS studies two complementary fingerprints of writing style: CH-SVM, a character-histogram classifier based on one-character marginal distributions, and FFT12-LR, a positional-signal classifier that represents selected characters and punctuation classes as impulse trains (binary indicator sequences over character positions) and extracts Fourier/Welch spectral descriptors. We also report CHiPS-F, a leakage-safe decision-level fusion variant, and an optional top-5 listwise reranker trained only on out-of-fold predictions. The method requires no tokenization, syntactic analysis, pretrained language model, or transformer fine-tuning, and it avoids character n-gram features with n ≥ 2 in the histogram component. On a locked grouped ROST split comprising 400 files from 392 source-text groups, written by 10 authors, with source-text-level evaluation and grouped five-fold model selection, CHiPS-F reaches 0.9310 accuracy and 0.9341 macro-F1. A matched but unrestricted character 2--5-gram TF--IDF SVM comparator reaches 1.0000 accuracy and macro-F1 on the same held-out groups, so the contribution is not a claim of best possible classification accuracy. Instead, the experiments ask how far restricted, transparent character evidence can go under strict leakage control. On ROSTories-cleaned, a secondary ROST-overlapping corpus comprising 1,248 files from 1,240 source-text groups, written by 19 authors, the same protocol gives 0.8919 accuracy and 0.8708 macro-F1 for CHiPS-R.