MACHINE LEARNING AND DEPLOYMENTS (My Simple Use case)
1. INTRODUCTION
Machine learning is simply a form of AI that enables a system to learn from data rather than through explicit programming.
ML has been deployed over the years on various applications like: Social Media Recommender engines; In most popular deep learning and computer vision applications e.g self driving cars; in chatbots assistant like Siri; in Stock Market Prediction and so on.
There are many Machine Learning deployment techniques, some include:
(i) Most popular: save model after training with pickle, TFlite, keras model. e.t.c , then use saved model on web or mobile for whatever ML tasks
(ii) Real time model training when running on the web or mobile app tasks: this can be computationally expensive
Readmore about are Machine Learning deployment techniques: https://www.kdnuggets.com/2019/06/approaches-deploying-machine-learning-production.html
2. PROJECT
This project demonstrates A simple supervised learning Multiple Logistic Regression Model used in a Classification of animal. It is deployed on a web application built with Python Django and other technologies, and the second deployment technique is used.
Link to app: https://animalpred.herokuapp.com/animalpre/
2.1 FRONTEND OF APPLICATION (webapp)
How the prediction works:
Once you answer all the questions and click predict Animal Class:
(i) It stores all your reply in a pandas dataframe .i.e a dictionary
(ii) Training of the ML model with Logistic regression classifier algorithm begins, using a Training data stored on a cloud storage.
(iii) it then feeds the stored reply into the trained ML model for prediction.
(iv) It predicts the class of the animal based on animal features you filled in
2.2 BACKEND OF APPLICATION (Code)
3. ADDED FUNCTIONALITIES OF ANIMAL PREDICT WEB APPLICATION:
(i) Stores every prediction as an API:
AFTER SIGNING UP
(ii) Provides latest updates about animal all over the world using a news api
(iii) Lets user search for any animal in mind
(iv) Upcoming features:
4. CONTRIBUTIONS OF PROJECT:
(i) Improves Animal datasets worldwide
(ii) Provides animal lovers latest animal news
(iii) More knowledge about animals
5. LINKS:
ANIMALPRE APP: https://animalpred.herokuapp.com/animalpre/
EXPLANATORY NOTEBOOK FOR THE ML MODEL: https://github.com/juniorboycoder/ANIMAL-PREDICT-ML-MODEL-USED-IN-THE-ANIMALPRE-WEBAPP/blob/master/ANIMALPREDICT.ipynb
GITHUB CODE: https://github.com/juniorboycoder/ANIMALPRE-DEPLOYED
TAGS: #MachineLearning #ArtificialIntelligence #Python #Django #Datascience
MY SOCIAL MEDIA
Twitter: https://twitter.com/oluyedejun1
Linkedln: https://www.linkedin.com/in/oluyede-segun-adedeji-jr-a5550b167/
