Machine learning has taken time to move into the space of academic economics. This is because empirical research in economics is concentrated on the identification of causal relationships in parsimonious statistical models; whereas machine learning is oriented towards prediction and is generally uninterested in either causality or parsimony. That leaves a gap for students, academics, and professionals who lack a standard reference on machine learning for economics and finance.
This book focuses on economic and financial problems with an empirical dimension, where machine learning methods may offer something of value. This includes coverage of a variety of discriminative deep learning models (DNNs, CNNs, LSTMs, and DQNs), generative machine learning models (GANs and VAEs), and tree-based models. It also covers the intersection of empirical methods in economics and machine learning, including regression analysis, natural language processing, and dimensionality reduction.
TensorFlow offers a toolset that can be used to define and solve any graph-based model, including those commonly used in economics. This book is structured to teach through a sequence of complete examples, each framed in terms of a specific economic problem of interest or topic. This simplifies otherwise complicated concepts, enabling the reader to solve workhorse theoretical models in economics and finance using TensorFlow.
What You'll Learn• Define, train, and evaluate machine learning models in TensorFlow 2• Apply fundamental concepts in machine learning, such as deep learning and natural language processing, to economic and financial problems • Solve theoretical models in economics

Who This Book Is ForStudents, data scientists working in economics and finance, public and private sector economists, and academic social scientists

机器学习需要时间来进入学术经济学的空间。这是因为经济学的实证研究主要集中在识别准统计模型中的因果关系；而机器学习则以预测为导向，通常对因果关系或准统计不感兴趣。这给学生、学者和专业人士留下了空白，他们缺乏关于经济和金融的机器学习的标准参考资料。

本书重点关注具有实证性的经济和金融问题，在这些问题上，机器学习方法可能提供一些价值。这包括对各种判别性深度学习模型（DNNs、CNNs、LSTMs和DQNs）、生成性机器学习模型（GANs和VAEs）和基于树的模型的报道。它还涵盖了经济学和机器学习的经验方法的交叉点，包括回归分析、自然语言处理和降维。

TensorFlow提供了一个工具集，可以用来定义和解决任何基于图的模型，包括经济学中常用的模型。本书的结构是通过一连串完整的例子进行教学，每个例子都是以一个感兴趣的具体经济问题或主题为框架。这简化了其他复杂的概念，使读者能够使用TensorFlow解决经济学和金融学的主要理论模型。

你将学会--在TensorFlow中定义、训练和评估机器学习模型 2--将机器学习的基本概念，如深度学习和自然语言处理，应用于经济和金融问题 --解决经济学的理论模型

本书适用对象学生、从事经济和金融工作的数据科学家、公共和私营部门的经济学家以及学术社会科学家