Название: Optimized Predictive Models in Health Care Using Machine Learning
Автор: Sandeep Kumar, Anuj Sharma, Navneet Kaur, Lokesh Pawar
Издательство: Wiley-Scrivener
Год: 2024
Страниц: 385
Язык: английский
Формат: pdf (true)
Размер: 34.4 MB

This book is a comprehensive guide to developing and implementing optimized predictive models in healthcare using Machine Learning and is a required resource for researchers, healthcare professionals, and students who wish to know more about real-time applications.

The book focuses on how humans and computers interact to ever-increasing levels of complexity and simplicity and provides content on the theory of optimized predictive model design, evaluation, and user diversity. Predictive modeling, a field of Machine Learning, has emerged as a powerful tool in healthcare for identifying high-risk patients, predicting disease progression, and optimizing treatment plans. By leveraging data from various sources, predictive models can help healthcare providers make informed decisions, resulting in better patient outcomes and reduced costs.

Deep Learning (DL) algorithms are crucial for enterprises to make business choices based on predictions. DL performs automatic data-oriented feature extraction and learns representations from raw data, unlike conventional ML algorithms. DL models use several neural network-based processing layers to understand illustrations of data with different levels of abstraction. Based on the data it gets as inputs from the layer below, each layer of a DL system constructs a representation of the observed patterns by maximizing local unsupervised criteria. The critical component of DL is that these layers of characteristics are found from data using a general-purpose learning technique rather than being built by human engineers. DL has demonstrated impressive performances in object detection in speech recognition, object detection in images, and natural language understanding. DL is particularly effective at identifying complex structures in high-dimensional data.

As the ML and DL fields are growing continuously, the challenge of training algorithms on large datasets has become increasingly complex. This challenge has led to the development of distributed machine learning, which allows data processing across multiple platforms. One of the newest extensions of this technology is Federated Learning (FL), which takes distributed Machine Learning a step further by enabling the training of
ML and DL algorithms using data stored on various devices, including computers, smartphones, and other mobile devices. This approach reduces the complexity of data administration and storage by transferring the calculation to the location where the data are produced. It necessitates that the feature spaces that each participating node shares be identical, which improves the delay and delivers greater security and privacy. Additionally, FL offers the advantage of providing improved tailored recommendations quickly to customers, with applications in areas such as movies, restaurants, and healthcare. FL has the advantage of transferring the computation to the data generation site, thereby reducing the complexity of data storage and administration. This is particularly useful for organizations with large datasets, as it enables them to train their algorithms without worrying about the storage and maintenance of massive data sets.

Other essential features of the book include:
provides detailed guidance on data collection and preprocessing, emphasizing the importance of collecting accurate and reliable data;
explains how to transform raw data into meaningful features that can be used to improve the accuracy of predictive models;
gives a detailed overview of Machine Learning algorithms for predictive modeling in healthcare, discussing the pros and cons of different algorithms and how to choose the best one for a specific application;
emphasizes validating and evaluating predictive models;
provides a comprehensive overview of validation and evaluation techniques and how to evaluate the performance of predictive models using a range of metrics;
discusses the challenges and limitations of predictive modeling in healthcare;
highlights the ethical and legal considerations that must be considered when developing predictive models and the potential biases that can arise in those models.

Audience:
The book will be read by a wide range of professionals who are involved in healthcare, Data Science, and Machine Learning.