iconcern - Intelligent College-Based Student Concern Classification and Routing System

A complete web-based system for NWSSU that uses SVM (Support Vector Machine) machine learning to automatically classify and route student concerns to the appropriate college or department.

Features

• Student Portal: Submit concerns with automatic AI classification
• Admin/Staff Dashboard: Manage concerns, update status, view routing
• SVM Classification: Automatic categorization of concerns (Academic, Financial, Guidance, Facilities, IT, Library)
• Intelligent Routing: Auto-route concerns to appropriate departments
• Status Tracking: Track concern status (Pending → In Progress → Resolved)
• Notifications: Real-time notifications for status updates
• Reports & Analytics: View statistics and insights

Technology Stack

• Backend: PHP 7.4+
• Database: MySQL (via XAMPP)
• Frontend: HTML5, CSS3, JavaScript
• Machine Learning: Python 3.x with scikit-learn (SVM)
• Server: XAMPP (Apache + MySQL)

Installation & Setup

Quick Setup (Recommended)

Option 1: Automated Setup (Easiest)

1. Ensure XAMPP is running (Apache and MySQL)
2. Open your browser and navigate to:

   http://localhost/iconcern/setup.php
   

3. Follow the on-screen instructions to:
   • Initialize the database
   • Train the SVM model with 1000 entries
4. Login with admin credentials (username: admin, password: admin123)

Option 2: Manual Setup

Prerequisites

1. XAMPP installed and running

   • Download from: https://www.apachefriends.org/
   • Ensure Apache and MySQL are running

2. Python 3.x with required packages

   pip install scikit-learn pandas numpy

Step 1: Clone/Copy Project

1. Copy the iconcern folder to your XAMPP htdocs directory:

   C:\xampp\htdocs\iconcern\
   

Step 2: Database Setup

Method A: Using PHP Script (Recommended)

php database/init_database.php

Method B: Using phpMyAdmin

1. Open phpMyAdmin: http://localhost/phpmyadmin
2. Import the database schema:
   • Click "Import" tab
   • Choose file: database/schema.sql
   • Click "Go"

The database iconcern_db will be created with all required tables, including:

• All 6 NWSSU colleges (COED, CCIS, CCJS, COM, CEA, CON)
• Complete department list
• Admin account (username: admin, password: admin123)

Step 3: Configure Database Connection

Edit config/database.php if your MySQL credentials are different:

define('DB_HOST', 'localhost');
define('DB_USER', 'root');
define('DB_PASS', '');  // Your MySQL password
define('DB_NAME', 'iconcern_db');

Step 4: Train the SVM Model

The system includes a comprehensive 1000-entry training dataset for high accuracy classification.

1. Open terminal/command prompt
2. Navigate to the project directory:

   cd C:\xampp\htdocs\iconcern

3. Run the training script:

   python ml/train_model.py

This will:

• Load 1000 training entries from ml/training_data.csv
• Train the SVM model with 97%+ accuracy
• Create svm_model.pkl and vectorizer.pkl files in the ml directory

Expected Output:

Model Accuracy: 97.50%
Training completed successfully!

Step 6: Access the Application

1. Start XAMPP (Apache and MySQL)
2. Open your browser and navigate to:

   http://localhost/iconcern/
   

3. You will be redirected to the login page

Default Login Credentials

Admin Account

• Username: admin
• Password: admin123

Student Account

• Username: student1
• Password: admin123

Staff Account

• Username: staff1
• Password: admin123

⚠️ Important: Change these default passwords in production!

Project Structure

iconcern/
├── admin/                  # Admin panel pages
│   ├── concerns.php
│   ├── routing.php
│   ├── users.php
│   └── reports.php
├── assets/                 # Static assets
│   ├── css/
│   │   └── style.css
│   └── js/
│       └── main.js
├── config/                 # Configuration files
│   ├── config.php
│   └── database.php
├── database/               # Database schema
│   └── schema.sql
├── includes/               # PHP includes
│   ├── auth.php
│   ├── concern.php
│   ├── college.php
│   ├── header.php
│   └── footer.php
├── ml/                     # Machine Learning scripts
│   ├── classify.py
│   └── train_model.py
├── uploads/                # File uploads directory
├── index.php              # Main dashboard
├── login.php              # Login page
├── register.php           # Registration page
├── submit_concern.php     # Submit concern form
├── my_concerns.php       # Student concerns list
├── concern.php           # Concern details page
├── notifications.php     # Notifications page
├── profile.php           # User profile
├── logout.php            # Logout handler
└── README.md             # This file

Database Schema

The system uses 9 main tables:

1. users - User accounts (students, staff, admins)
2. colleges - NWSSU colleges (COED, CCIS, CCJS, COM, CEA, CON)
3. departments - Administrative departments (Guidance, Cashiers, Registrar, Library, OSAS, IT Support, Maintenance, and all colleges as departments)
4. concerns - Student-submitted concerns
5. classifications - SVM classification results with confidence scores
6. routing - Concern routing information (college and department)
7. status_history - Status change logs with notes
8. notifications - User notifications
9. training_data - ML training data (1000 entries included)

How It Works

1. Student Submits Concern

• Student logs in and fills out the concern form
• Optionally uploads an attachment (image/document)

2. Automatic Classification

• System calls Python SVM script (ml/classify.py)
• Concern text is preprocessed and classified
• Result is saved in classifications table

3. Intelligent Classification & Routing

• SVM Classification: Concern is automatically classified into categories:
  • Academic, Financial, Guidance, Facilities, IT, Library, General
• College Detection: System detects college names in concern text (e.g., "CCIS", "COED", "CEA")
• Smart Routing: Based on classification AND detected college:
  • Academic → Registrar Office (or specific college if detected)
  • Financial → Cashiers Office
  • Guidance → Guidance Office
  • Facilities → Maintenance Office (or specific college department if college detected)
  • IT → IT Support
  • Library → Library Office

Example: "The TV in the CCIS classroom is broken" → Classified as Facilities → Routed to CCIS Department

4. Staff/Admin Management

• Staff can view routed concerns
• Update status (Pending → In Progress → Resolved)
• Add notes and track history

5. Notifications

• Students receive notifications when status changes
• Real-time updates via notification system

SVM Classification

The system uses a Support Vector Machine (SVM) classifier trained on 1000 labeled entries with:

• Text preprocessing: Lowercasing, special character removal, whitespace normalization
• TF-IDF Vectorization: 2000 features with n-grams (1-3) for better accuracy
• Linear SVM: Fast and accurate for text classification
• Model Accuracy: 97.50% on test set
• Fallback classification: Keyword-based if model not available
• College Detection: Automatically detects college names (COED, CCIS, CCJS, COM, CEA, CON) in concern text

Categories

• Academic (200 training samples) - Course, grades, exams, enrollment
• Financial (150 training samples) - Tuition, fees, scholarships, payments
• Guidance (150 training samples) - Counseling, mental health, personal issues
• Facilities (200 training samples) - Buildings, maintenance, equipment
• IT (150 training samples) - Computers, network, software, technical issues
• Library (100 training samples) - Books, research, study rooms
• General (50 training samples) - General inquiries and other concerns

Customization

Adding New Colleges

Edit database/schema.sql and add to colleges table, or use phpMyAdmin to insert new records.

Adding New Departments

Insert into departments table via phpMyAdmin or SQL.

Modifying Routing Rules

Edit includes/concern.php in the autoRoute() method to change routing logic.

Training with Custom Data

1. Update ml/training_data.csv with your labeled data (format: concern_text,label)
2. Or add data to the training_data table in the database
3. Run training script: python ml/train_model.py
4. Model will be saved automatically

Note: The system includes 1000 pre-labeled training entries. You can add more to improve accuracy.

Troubleshooting

Database Connection Error

• Check XAMPP MySQL is running
• Verify credentials in config/database.php
• Ensure database iconcern_db exists

Python Classification Not Working

• Verify Python is installed: python --version
• Install required packages: pip install scikit-learn pandas numpy
• Check Python path in includes/concern.php (line with python command)
• On Windows, you may need to use python instead of python3

File Upload Issues

• Ensure uploads/ directory exists and is writable
• Check file size limits in config/config.php
• Verify allowed file extensions

Session Issues

• Ensure PHP sessions are enabled
• Check session directory permissions
• Clear browser cookies if login persists

Security Notes

⚠️ For Production Use:

1. Change all default passwords
2. Use strong password hashing (already using bcrypt)
3. Enable HTTPS/SSL
4. Sanitize all user inputs (already implemented)
5. Implement CSRF protection
6. Set proper file upload permissions
7. Hide sensitive error messages
8. Use environment variables for database credentials

Support

For issues or questions:

• Check the database connection settings
• Verify all dependencies are installed
• Review Apache error logs: C:\xampp\apache\logs\error.log
• Check PHP error logs

License

This project is developed for NWSSU (Northwest Samar State University).

Version

Current Version: 1.0.0

---

Developed for NWSSU | Intelligent Concern Classification System