http://Ted-recommender.com Web Application built for reseach purpose. The mission is to apply basic concept on Data Mining and Machine Learning to develop a simple search, classification, and recommendation system.

The whole web application consists of 3 features search, classification, and recommendation on a dataset of Ted-Talks videos from beginning to September 2017.
All there features are developed inside ted_engine.py file, which is initialized and used by ted.py file. This file is a flask application, that connects the logic from back-end to the front-end. All files inside folder static and templates are for the front-end HTML and CSS file.

To replication and got the application up running in localhost, there are just some simple steps next to be done.
After retrieving the project to a local machine

Step 1: Prepare the run environment:

Because the application is written using python 3.7.0 with few necessary libraries.
Make sure local machine have already installed python 3, flask, numpy, pandas, and nltk libraries.
If not you can easily install python 3 and pip tool to install the needed libraries (https://realpython.com/installing-python/).

Step 2: Execute the application:

Before run, make sure you change the app.run value from host="0.0.0.0", port="80" to nothing as default. The host and port are only for when you run the application on a host.
Simply go to the ted-recommender diectory
Run the command inside your terminal: python ted.py
Your application should be up running at http://localhost:5000
You can use your browser to go to this address and start using the application.

There is a search field for finding your favorite Ted-talks. There is also a recommend button under each result's videos.
Finally, the classification feature helps you to classify any Ted-talks contents that you can think of.
The mission is to apply basic concept on Data Mining and Machine Learning to develop a simple search, classification, and recommendation system.

For more details about how these features are implemented, please go to my homepage: https://tungpv.com

Reference:
Dataset: https://www.kaggle.com/rounakbanik/ted-talks

Search feature built base on: https://nbviewer.jupyter.org/url/crystal.uta.edu/~cli/cse5334/ipythonnotebook/P1.ipynb

Text classification and Naïve Bayes (https://nlp.stanford.edu/IR-book/pdf/13bayes.pdf).

Recommender System built base on: Chapter 9 Recommendation Systems, Mining of Massive Datasets by Jure Leskovec, Anand Rajaraman, Jeff Ullman