�

PROJECT TITLE:REVOLUTIONIZING CAREER DEVELOPMENT WITH-PERSONALIZED COURSE RECOMMENDATIONS

​

​

​

PROBLEM STATEMENT

In the era of digital learning, the vast and ever-expanding landscape of online courses presents a significant challenge for learners how to efficiently discover and access the most relevant educational content. Many learners struggle with choice overload, lack of personalized guidance, and the time-consuming process of sifting through numerous course options. To address these challenges, our project aims to develop a Course Recommendation System.

OBJECTIVE

The project's core objective is to build a Course Recommendation System that employs machine learning to offer personalized course suggestions to users. It aims to streamline the process of discovering online courses by filtering and presenting the most relevant options from vast databases. The system's algorithms will prioritize accuracy and relevance in recommendations, adapting to users' unique preferences and learning goals.

AIM AND SCOPE OF THE PROJECT

The system will employ machine learning to offer personalized course recommendations based on users' preferences and goals. It will provide accurate and relevant suggestions from vast course databases like Coursera .The user- friendly web application will offer an intuitive interface, allowing users to explore, access, and engage with course recommendations. Scalability will be a key consideration to accommodate a growing volume of courses.

• The platform proposes educational programs that cater to the specific needs and objectives of the learners, taking into account their competencies, ambitions, and favored learning methods.
• It encompasses a user-friendly interface that enables browsing and selecting courses, monitoring advancement, and enhancing the algorithm.

​

LITERATURE REVIEW

• P. S. Patel, J. Undavia, and D. Bhatti, “Master Course Selection Prediction Model Using Modify Hybrid NeuroFuzzy Inference System,” ICTACT J. Soft Comput., vol. 11, no. 01, pp. 2205– 2212, 2020, doi: 10.21917/ijsc.2020.0314.
• Deepjyoti Roy & Mala Dutta , “A systematic review and research perspective on recommender systems”, Journal of Big Data volume 9, Article number: 59 (2022)
• David Estrela, Sergio Batista & Diogo Martinho "A Recommendation System for Online Courses" Conference: World Conference on Information Systems and Technologies 

​

ARCHITECTURE DIAGRAM

UML DIAGRAM

METHODOLOGY

• This app is designed specifically to save time for young people who aspire to be in a greater position in future, My app uses ML to comprehend a persons need by analyzing the future position the user wants to reach and recommends best suited course to place in that track.
• There are various benefits in this app: 
• Personalized Dashboard
• Interactive workflow
• Databases with top online course providers
• User Friendly interface

MODULES

• ML model

​

• User Database

​

• Recommender System Integration

​

• Dashboard

• ML model: 
• This Module is the most important part of our project, this module gets the courses from the provided database that is synced with the application and recommends the best course for the given future path 

​

• User Database:
• This module is used to store the user database that is required for the ML model to recommend, and is implemented by firebase. This part has a special significance in the project as it is a vital part of the project without user database there is no personalized recommendations. 

​

• Recommender system integration: 
• The whole recommender system is integrated and implemented with the project in this module with the help of flask. 

​

• Dashboard: 
• The user dashboard displays the recommendations from poplar e learning websites personalized to the specific user. This is done with help of external html, css files that is linked with main code using flask library. 

​

​

1. Collaborative Filtering: Collaborative filtering methods, such as user-based and item-based filtering, can be employed to make recommendations based on the preferences and behavior of similar users or courses. 

​

2. Content-Based Filtering: This method considers the attributes and content of courses and matches them with a user's profile. It recommends courses that are similar in content to what the user has shown interest in. 

​

3. Matrix Factorization: Techniques like Singular Value Decomposition (SVD) and matrix factorization can help in identifying latent factors and making recommendations based on those factors. 

​

4. Machine Learning Algorithms: Various machine learning algorithms, including decision trees, random forests, and gradient boosting, can be used to create predictive models for course recommendations. 

​

5. Natural Language Processing (NLP): NLP techniques can be applied to analyze course descriptions and user feedback to extract relevant information and enhance recommendation accuracy. 

​

6. Clustering: Clustering methods like k-means can group similar courses together, aiding in providing recommendations based on course clusters. 

​

7. Hybrid Approaches: Combining multiple recommendation techniques, such as collaborative filtering and content-based filtering, can result in more accurate and diverse recommendations. 

​

8. Deep Learning: Deep neural networks, particularly in the form of deep recommender systems, can be used to model complex patterns and relationships in course data. 

​

9. Evaluation Metrics: Methods for evaluating the performance of the recommendation system, such as precision, recall, and mean average precision, can be implemented to ensure the quality of recommendations. 

​

10. Data Preprocessing: Techniques for cleaning and preparing data, handling missing values, and reducing noise in the dataset are essential for accurate recommendations. 

​

11. User Profiling: Creating and updating user profiles based on user interactions and feedback is crucial for understanding user preferences.

ALGORITHM: 

​

• DATA COLLECTION: 
• This module is responsible for collecting the course details from eLearning websites by web scraping methods. 
• Step 1: Find reliable sites to extract course details 
• Step 2: Clean the data and combine the cleaned datasets from various sources and pass as database for machine learning. 

​

• Machine Learning: 
• This module is used to parse the course details and make a machine learning model to be implemented in out web application. 
• Step 1: Access the cleaned and combined database with a python notebook file for efficient build of machine learning. 
• Step 2: Build the model and save as a .pkl extension which can be used in main.py for recommender using the tags built by the model by passing pickle file. 

​

​

HOMESCREEN �

COURSE RECOMMENDER ( LOGIN PAGE)

DASHBOARD

RECOMMENDATION BASED ON THE COURSE SELECTED

PROFILE SETUP PAGE FOR CREATING AN ACCOUNT

EXPLORE SCREEN IN THE WEBSITE