Data Collection and Analysis of French Dialects

TL;DR

This paper presents the creation and analysis of a new French dialect dataset, applying machine learning classifiers to classify dialect samples and evaluating data mining techniques within the CRISP-DM framework.

Contribution

It introduces a new French dialect dataset and evaluates machine learning classifiers for dialect classification using a structured data mining approach.

Findings

Effective classifiers identified for dialect classification

Key features for distinguishing dialects determined

Data quality issues addressed and mitigated

Abstract

This paper discusses creating and analysing a new dataset for data mining and text analytics research, contributing to a joint Leeds University research project for the Corpus of National Dialects. This report investigates machine learning classifiers to classify samples of French dialect text across various French-speaking countries. Following the steps of the CRISP-DM methodology, this report explores the data collection process, data quality issues and data conversion for text analysis. Finally, after applying suitable data mining techniques, the evaluation methods, best overall features and classifiers and conclusions are discussed.