{"id":20270394,"url":"https://github.com/ziadelsayed0/service-quality-ai-based-system","last_synced_at":"2026-05-13T02:05:28.846Z","repository":{"id":187630912,"uuid":"669892746","full_name":"ziadelsayed0/Service-Quality-Ai-based-System","owner":"ziadelsayed0","description":"Ai Application For Enhancing Customer Services","archived":false,"fork":false,"pushed_at":"2025-02-08T21:55:15.000Z","size":1102,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-02-08T22:26:56.346Z","etag":null,"topics":["arbert","bert","django-rest-framework","fine-tuning","marbert","sentiment-analysis","text-classification","topic-classification","topic-modeling"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ziadelsayed0.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-07-23T19:11:58.000Z","updated_at":"2025-02-08T21:55:18.000Z","dependencies_parsed_at":"2024-09-11T10:10:54.835Z","dependency_job_id":null,"html_url":"https://github.com/ziadelsayed0/Service-Quality-Ai-based-System","commit_stats":null,"previous_names":["ziadelsayed0/service-quality-ai-based-system"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ziadelsayed0%2FService-Quality-Ai-based-System","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ziadelsayed0%2FService-Quality-Ai-based-System/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ziadelsayed0%2FService-Quality-Ai-based-System/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ziadelsayed0%2FService-Quality-Ai-based-System/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ziadelsayed0","download_url":"https://codeload.github.com/ziadelsayed0/Service-Quality-Ai-based-System/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":241753145,"owners_count":20014252,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["arbert","bert","django-rest-framework","fine-tuning","marbert","sentiment-analysis","text-classification","topic-classification","topic-modeling"],"created_at":"2024-11-14T12:30:12.797Z","updated_at":"2026-05-13T02:05:23.801Z","avatar_url":"https://github.com/ziadelsayed0.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Service Quality AI-Based System\n\n## Introduction  \n\u003cimg src=\"img1.jpeg\" alt=\"drawing\" width=\"35%\" height=\"35%\" align=\"right\"/\u003e\n\nThe **Service Quality AI-Based System** is a state-of-the-art platform developed to revolutionize customer service and support, with a focus on Arabic-language industries. It achieves this through advanced natural language processing (NLP) techniques tailored to analyze customer interactions and generate solutions efficiently.  \n\n### Key Features :\n- **Topic Classification**: Automatically categorizes customer queries into predefined topics (F1 Score: 95.8%).  \n- **Sentiment Analysis**: Identifies the emotional tone of customer interactions (F1 Score: 96%).  \n- **Solution Generation**: Provides actionable responses to customer queries.  \n\n### Technologies Used :\n- **Programming Language**: Python  \n- **Framework**: Django  \n- **AI Models**: MARBERT, ARBERT, GPT, MT5  \n- **Tools**: Web scraping tools, tokenization libraries, hyperparameter tuning methods  \n\n---\n\n## Project Phases  \n\n### 1. Data Gathering  \nData is collected using web scraping tools to ensure a comprehensive dataset covering diverse scenarios and linguistic nuances in Arabic. This supports robust training for the AI models.  \n\n### 2. Data Preprocessing  \nTo prepare the data for training, it undergoes the following steps:  \n- **Noise Removal**: Cleaning special characters and irrelevant text.  \n- **Tokenization and Normalization**: Standardizing Arabic language data.  \n- **Data Splitting**: Dividing data into training, validation, and test sets.  \n\n### 3. Hyperparameter Tuning  \nHyperparameter optimization ensures the models perform optimally. This involves fine-tuning configurations such as learning rates, batch sizes, and architecture parameters.  \n\n### 4. AI Models Utilized  \n| **Task**                | **Model**   | **Dataset**                  | **F1 Score** |\n|--------------------------|-------------|------------------------------|--------------|\n| **Topic Classification** | MARBERT     | Merged Topics dataset    | 95.8%        |\n| **Sentiment Analysis**    | ARBERT      | 100k revs \u0026 glare dataset     | 96.0%        |\n| **Solution Generation**   | GPT, MT5    | Contextual solution dataset  | N/A          |\n\n### 5. Deployment  \nThe system is deployed using a robust **Django framework**, providing:  \n- A seamless **frontend** and **backend** experience.  \n- A responsive and user-friendly interface for end-users.\n\n  \n\u003ctable\u003e\n  \u003ctr\u003e\n    \u003ctd\u003e\u003cimg src=\"Home.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n    \u003ctd\u003e\u003cimg src=\"Log-in.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n    \u003ctd\u003e\u003cimg src=\"Registration.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003e\u003cimg src=\"Complaint.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n    \u003ctd\u003e\u003cimg src=\"Adminstration.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n    \u003ctd\u003e\u003cimg src=\"About-team.jpg\" width=400 hieght=400/\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \n\u003c/table\u003e\n\n---\n\n## How It Works  \n1. **Data Input**: Customer queries are ingested into the system.  \n2. **Processing**: The platform performs topic classification, sentiment analysis, and solution generation.  \n3. **Output**: Contextual responses or solutions are provided in real time.\n\n---\n\n## Colab Notebooks  \n\nExplore the Colab notebooks for:  \n- Topic Classification Notebook [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1Kyux7JPYJEbJKY3LjNZDhyezXChUQV0p?usp=sharing) \n\n*Click the links above to access the respective Colab notebooks and view the implementation details.*\n\n---\n\n## Benefits  \n- Accurate and efficient processing of Arabic-language customer interactions.  \n- Scalable and reliable system design.  \n- Simplifies customer support and enhances service quality.  \n\n---\n\n## About the Models:\n### ARBERT \u0026 MARBERT: Deep Bidirectional Transformers for Arabic\n\u003cimg src=\"ARBERT_MARBERT.jpg\" alt=\"drawing\" width=\"30%\" height=\"30%\" align=\"right\"/\u003e\n\n* introduce ```ARBERT``` and ```MARBERT```, two powerful Transformer-based language models for Arabic;\n* introduce ```ArBench```, a multi-domain, multi-variety benchmark for Arabic naturaal language understanding based on 41 datasets across 5 different tasks and task clusters;\n* evaluate ARBERT and MARBERT on ArBench and compare against available language models.\n\nModel establish new state-of-the-art (SOTA) on all 5 tasks/task clusters on 37 out of the 41 datasets.\nLanguage models are publicaly available for research (see below).\nThe rest of this repository provides more information about the language models, benchmark, and experiments.\n### How to use ARBERT and MARBERT\n\n#### Loading directly from Huggingface\nYou can use ARBERT and MARBERT with [Hugging Face's Transformers](https://github.com/huggingface/transformers) library as follow.\n \n ```python\n    from transformers import AutoTokenizer, AutoModel\n    #load AEBERT model from huggingface\n    ARBERT_tokenizer = AutoTokenizer.from_pretrained(\"UBC-NLP/ARBERT\")\n    ARBERT_model = AutoModel.from_pretrained(\"UBC-NLP/ARBERT\")\n  \n    #load MAEBERT model from huggingface\n    MARBERT_tokenizer = AutoTokenizer.from_pretrained(\"UBC-NLP/MARBERT\")\n    MARBERT_model = AutoModel.from_pretrained(\"UBC-NLP/MARBERT\") \n ```\n### 4.2 Example: Fine-tuning MARBERT for Sentiment Analysis\nMARBERT Fine-Tuning demo with PyTorch checkpoint for Sentiment Analysis on the AJGT dataset [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1M0ls7EPUi1dwqIDh6HNfJ5y826XvcgGX?usp=sharing)\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fziadelsayed0%2Fservice-quality-ai-based-system","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fziadelsayed0%2Fservice-quality-ai-based-system","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fziadelsayed0%2Fservice-quality-ai-based-system/lists"}