

Evaluation Of Public Opinion From Nutrition
Feedback Through NLP
Student Name:Shiran Azulay
Advisor: Mr. Yoram Rozenwaks
Software Engineering
Algorithms:
Algorithm of scrapping and nlp predicts:
These models will predict people opinion about health and diets using nlp
• Html of pages on web in health and diets.
• Pkl files – one for classifier the sentiment for the graphs and one for stop words
•
There are 3 links , each link has is own csv with summary of the comments
•
Train the data( divide to train and test) and get accuracy above 85%
•
Scrapping the data to get author and title of the article using Beautiful Soup
•
Get the subject by frequency keywords of health and diets using sentiment and nlp algorithm
An example of the results:
My project is all about health and diets in machine learning
world, to give a relevant information for dietitians ,fitness
trainers, companies and for the general public .I analyzed the
comments of defines 3 websites using nlp and predict the
details of the(Author, title and subject).the results is shown
in web page that open to the open audience.
System Architectural Design:
Scrapping part:
•each website has its own comments, scrap the comments using ML model.
•The name of the author and the title of the website we scarp using html.
Predication with NLP – provide services of:
•Database – store all the comments and the result of nlp into csv files
•Graphs-the nlp algorithm will predict the opinion of the comment and will count the
number of positive, negative, not defined and the nlp didn’t succeed to predict that.
•Matching by AI –remove stop words from text in preprocessing using the pkl file.
•Matching sentiment by AI– using the classifier.pkl for predict the sentiment of the
comment, uses a data set of positive and negative words.
Web page :
•Home- will present the name of the projects and photos of health and diets
•Articles – present the websites for them we scarp the data
•Research Results-the result of the project
•About-in my vision I am telling about the project and I search about that
•Contact- my details to connect with me
User Interface Design:
The following design describes the web page as a blog .