Book Image

Hands-On Machine Learning with C++

By : Kirill Kolodiazhnyi
Book Image

Hands-On Machine Learning with C++

By: Kirill Kolodiazhnyi

Overview of this book

C++ can make your machine learning models run faster and more efficiently. This handy guide will help you learn the fundamentals of machine learning (ML), showing you how to use C++ libraries to get the most out of your data. This book makes machine learning with C++ for beginners easy with its example-based approach, demonstrating how to implement supervised and unsupervised ML algorithms through real-world examples. This book will get you hands-on with tuning and optimizing a model for different use cases, assisting you with model selection and the measurement of performance. You’ll cover techniques such as product recommendations, ensemble learning, and anomaly detection using modern C++ libraries such as PyTorch C++ API, Caffe2, Shogun, Shark-ML, mlpack, and dlib. Next, you’ll explore neural networks and deep learning using examples such as image classification and sentiment analysis, which will help you solve various problems. Later, you’ll learn how to handle production and deployment challenges on mobile and cloud platforms, before discovering how to export and import models using the ONNX format. By the end of this C++ book, you will have real-world machine learning and C++ knowledge, as well as the skills to use C++ to build powerful ML systems.
Table of Contents (19 chapters)
Section 1: Overview of Machine Learning
Section 2: Machine Learning Algorithms
Section 3: Advanced Examples
Section 4: Production and Deployment Challenges


In this chapter, we examined anomalies in data. We discussed several approaches to anomaly detection and looked at two kinds of anomalies: outliers and novelties. We considered the fact that anomaly detection is primarily an unsupervised learning problem, but despite this, some algorithms require labeled data, while others are semi-supervised. The reason for this is that, generally, there is a tiny number of positive examples (that is, anomalous samples) and a large number of negative examples (that is, standard samples) in anomaly detection tasks.

In other words, we usually don't have enough positive samples to train algorithms. That is why some solutions use labeled data to improve algorithm generalization and precision. On the contrary, supervised learning usually requires a large number of positive and negative examples, and their distribution needs to be balanced...