EgnumaAdane/Phonetic-Similarity-Tool

Custom Python tool for algorithmic analysis and structural similarity scoring across Amharic and Tigrigna using phoneme and syllable decomposition. # SimScore-Analyze Custom Python tool for algorithmic analysis and structural similarity scoring across Amharic and Tigrigna using phoneme and word level decomposition. --- ## 📜 Table of Contents - Project Overview - 📊 Key Findings - Phoneme Overlap - Word-Level Similarity - 🛠️ Technical Implementation - ⚙️ Requirements and Installation - 💡 Execution - 📄 License --- ## 🎯 Project Overview This project addresses a critical challenge in building language technology (specifically Text-to-Speech systems) for **under-resourced Ethiosemitic languages, Amharic and Tigrigna**. The primary objective was to **computationally quantify the structural similarity** between the two languages to provide a data-driven justification for the feasibility of resource reuse in application development. ## 📊 Key Findings: Data-Driven Resource Strategy The Python analysis provided a comprehensive view of the similarities and differences, revealing a high distinction between vocabulary and phonetic structure. ### Phoneme Overlap | Metric | Value | | :--- | :--- | | **Overall Phoneme Overlap** | **83.15%** | * **Impact:** This high degree of shared sound structure strongly supports the strategic decision to reuse phonetic models across both languages, drastically reducing engineering complexity and development time. ### Word-Level Similarity | Metric | Value | | :--- | :--- | | **Word Level Similarity** | **3.62%** | * **Impact:** Confirms that the two languages maintain distinct vocabularies, meaning shared development efforts must be concentrated at the sub-word (phonetic) level rather than the word level. ## 💻 Technical Implementation The analysis required the engineering of a specialized, multi-stage Python pipeline due to the lack of standard library support for the Ethiosemitic Ge'ez script. **Core Technical Components:** * **Custom Transliteration Mapping:** A dictionary-based data structure was implemented to map Ge'ez characters to a Latin-based phonetic represent …