🌐 Multilingual Translator & ML Model Evaluation

A multilingual Streamlit application combining text translation, transliteration, text-to-speech, voice-to-text, and interactive machine-learning model evaluation.

Overview • Features • Translation • ML Evaluation • Tech Stack • Getting Started

---

✨ Overview

Multilingual Translator & ML Model Evaluation is a Streamlit-based Python application that brings together two major capabilities in one interface:

🌍 Multilingual Language Tools

• Translate text into a selected target language
• Romanize / transliterate translated text
• Convert translated text into speech
• Capture speech and convert it into text

🤖 Machine Learning Evaluation

• Clustering with K-Means
• Regression with Linear Regression
• Classification with multiple supervised learning algorithms
• Data preprocessing and feature scaling
• Imbalanced-data handling with SMOTE
• Visual evaluation using plots, heatmaps, and confusion matrices

The result is a practical application combining NLP, speech processing, data preprocessing, machine learning, and interactive visualization.

---

🚀 Features

Feature	Description
🌐 Multilingual Translation	Translate text into multiple target languages
🔤 Romanization	Convert translated text into a more readable Romanized form
🔊 Text-to-Speech	Generate audio from translated text
🎙️ Voice-to-Text	Capture audio input and convert speech into text
🧩 Clustering	Explore K-Means clustering
📈 Regression	Train and evaluate Linear Regression models
🧠 Classification	Compare Logistic Regression, Naive Bayes, SVM, and KNN
⚖️ SMOTE	Handle imbalanced classification datasets
📏 Feature Scaling	Scale numerical features with StandardScaler
📊 Visualization	Generate correlation heatmaps, interactive plots, and confusion matrices

---

🌍 Translation Workflow

                 ┌────────────────────┐
                 │   User Text / Voice│
                 └─────────┬──────────┘
                           │
               ┌───────────┴───────────┐
               ▼                       ▼
       ┌───────────────┐       ┌───────────────┐
       │    Text Input │       │  Voice Input  │
       └───────┬───────┘       └───────┬───────┘
               │                       │
               │                       ▼
               │              Speech-to-Text
               │                       │
               └───────────┬───────────┘
                           ▼
                  GoogleTranslator
                           │
                           ▼
                 Translated Text
                           │
                  ┌────────┴────────┐
                  ▼                 ▼
            Transliteration    Text-to-Speech
                  │                 │
                  ▼                 ▼
          Romanized Output      Audio Output

---

🔤 Translation Capabilities

The translation workflow supports:

• 🌐 Source-language text input
• 🎯 Target-language selection
• 🔤 Romanization / transliteration
• 🔊 Audio generation through gTTS
• 🎙️ Voice-to-text input through speech_recognition

The translation component uses GoogleTranslator for language conversion and gTTS for text-to-speech generation.

---

🤖 Machine Learning Evaluation

The application also provides an interactive environment for evaluating several machine-learning approaches.

🧩 Clustering

K-Means

Used to group data points based on similarity.

Dataset
   │
   ▼
Feature Preparation
   │
   ▼
K-Means Clustering
   │
   ▼
Cluster Assignments
   │
   ▼
Interactive Visualization

---

📈 Regression

Linear Regression

Used for predicting continuous target variables.

The workflow includes:

• Data preparation
• Feature/target separation
• Model training
• Prediction
• Evaluation
• Visualization

---

🧠 Classification

The application supports several classification algorithms:

Model	Type
Logistic Regression	Linear classifier
Naive Bayes	Probabilistic classifier
Support Vector Machine	Margin-based classifier
KNeighborsClassifier	Instance-based classifier

The interface allows models to be trained and evaluated on prepared datasets.

---

🧹 Data Preprocessing

Before model evaluation, the application supports several preprocessing steps.

Missing Values

Missing values can be handled during dataset preparation.

Categorical Encoding

Categorical variables are encoded using:

LabelEncoder

Feature Scaling

Numerical features are scaled using:

StandardScaler

Class Balancing

For imbalanced classification datasets, the project applies:

SMOTE

This provides a more balanced training distribution before model evaluation.

---

📊 Visualization & Evaluation

The application includes visualization tools for understanding datasets and model performance.

Available Visualizations

• 🔥 Correlation heatmaps
• 📊 Interactive Plotly visualizations
• 🔲 Confusion matrices
• 📈 Model evaluation outputs
• 🧩 Clustering visualizations
• 📉 Regression evaluation views

---

🧠 Application Architecture

                 ┌─────────────────────┐
                 │     Streamlit UI    │
                 └──────────┬──────────┘
                            │
           ┌────────────────┼────────────────┐
           │                │                │
           ▼                ▼                ▼
     Language Tools     ML Workflow      Visualization
           │                │                │
     ┌─────┼─────┐      ┌───┼────┐           │
     │     │     │      │   │    │           │
     ▼     ▼     ▼      ▼   ▼    ▼           ▼
 Translate TTS  STT   Cluster Regr Class   Charts
     │                      │     │     │
     └──────────────────────┴─────┴─────┘
                    │
                    ▼
              Streamlit Output

---

🛠️ Tech Stack

---

📁 Repository Structure

Multiligual_translator/
│
├── 📂 images/
├── 🐍 Translator.py
├── 📦 requirements.txt
└── 📘 README.md

---

⚙️ Getting Started

1️⃣ Clone the Repository

git clone https://github.com/Divakar1326/Multiligual_translator.git
cd Multiligual_translator

2️⃣ Create a Virtual Environment

Windows

python -m venv .venv
.venv\Scripts�ctivate

macOS / Linux

python3 -m venv .venv
source .venv/bin/activate

3️⃣ Install Dependencies

pip install -r requirements.txt

4️⃣ Run the Streamlit Application

streamlit run Translator.py

---

🧪 Typical Usage

🌍 Translation

1. Enter text or provide voice input.
2. Select a target language.
3. Translate the input.
4. View the translated output.
5. Romanize the translation when available.
6. Generate audio from the translated result.

🤖 Model Evaluation

1. Provide or load a dataset.
2. Prepare the features.
3. Apply preprocessing.
4. Select clustering, regression, or classification.
5. Train and evaluate the selected model.
6. Inspect metrics and visualizations.

---

📦 Core Python Libraries

streamlit
numpy
pandas
matplotlib
seaborn
scikit-learn
imbalanced-learn
gTTS
speech-recognition
transliterate
plotly

The repository's requirements.txt should remain the final source of truth for the exact dependency versions.

---

🧠 What This Project Demonstrates

• 🌐 Multilingual NLP application development
• 🎙️ Speech-to-text integration
• 🔊 Text-to-speech generation
• 🔤 Transliteration / Romanization
• 🧹 Data preprocessing
• 📏 Feature scaling
• ⚖️ Imbalanced-data handling
• 🤖 Classification
• 📈 Regression
• 🧩 Clustering
• 📊 Interactive visualization
• 🖥️ Streamlit application development

---

🔮 Future Improvements

Potential extensions include:

• 🌐 Add more translation providers or offline translation models
• 💬 Add translation history
• 🎙️ Improve real-time voice interaction
• 🧠 Add transformer-based NLP models
• 📊 Add richer model-comparison dashboards
• 📈 Add cross-validation and additional evaluation metrics
• 🚀 Deploy the application publicly
• 🧪 Add automated tests and validation

---

⚠️ Notes

• Internet connectivity may be required for external translation and text-to-speech services.
• Model results depend on the dataset and preprocessing configuration.
• Voice recognition quality can vary depending on audio quality and environment.
• This project is intended for educational and demonstration purposes.

---

👨‍💻 Author

Divakar M

B.Tech CSE — Artificial Intelligence & Data Science

AI/ML • Generative AI • Python • NLP • Machine Learning

---

⭐ If you find this project useful, consider starring the repository.

Translate 🌍 • Analyze 🤖 • Visualize 📊