Book Image

Intelligent Projects Using Python

By : Santanu Pattanayak
Book Image

Intelligent Projects Using Python

By: Santanu Pattanayak

Overview of this book

This book will be a perfect companion if you want to build insightful projects from leading AI domains using Python. The book covers detailed implementation of projects from all the core disciplines of AI. We start by covering the basics of how to create smart systems using machine learning and deep learning techniques. You will assimilate various neural network architectures such as CNN, RNN, LSTM, to solve critical new world challenges. You will learn to train a model to detect diabetic retinopathy conditions in the human eye and create an intelligent system for performing a video-to-text translation. You will use the transfer learning technique in the healthcare domain and implement style transfer using GANs. Later you will learn to build AI-based recommendation systems, a mobile app for sentiment analysis and a powerful chatbot for carrying customer services. You will implement AI techniques in the cybersecurity domain to generate Captchas. Later you will train and build autonomous vehicles to self-drive using reinforcement learning. You will be using libraries from the Python ecosystem such as TensorFlow, Keras and more to bring the core aspects of machine learning, deep learning, and AI. By the end of this book, you will be skilled to build your own smart models for tackling any kind of AI problems without any hassle.
Table of Contents (12 chapters)

Taking class imbalances into account

Class imbalance is a major problem when it comes to classification. The following diagram depicts the class densities of the five severity classes:

Figure 2.4: Class densities of the five severity classes

As we can see from the preceding chart, nearly 73% of the training data belongs to Class 0, which stands for no diabetic retinopathy condition. So if we happen to label all data points as Class 0, then we would have 73% percent accuracy. This is not desirable in patient heath conditions. We would rather have a test say a patient has a certain heath condition when it doesn't (false positive) than have a test that misses detecting a certain heath condition when it does (false negative). A 73% accuracy may mean nothing if the model learns to classify all points as belonging to Class 0.

Detecting the higher severity classes are more important...