Book Image

Hands-On Data Science with Anaconda

By : Yuxing Yan, James Yan
Book Image

Hands-On Data Science with Anaconda

By: Yuxing Yan, James Yan

Overview of this book

Anaconda is an open source platform that brings together the best tools for data science professionals with more than 100 popular packages supporting Python, Scala, and R languages. Hands-On Data Science with Anaconda gets you started with Anaconda and demonstrates how you can use it to perform data science operations in the real world. The book begins with setting up the environment for Anaconda platform in order to make it accessible for tools and frameworks such as Jupyter, pandas, matplotlib, Python, R, Julia, and more. You’ll walk through package manager Conda, through which you can automatically manage all packages including cross-language dependencies, and work across Linux, macOS, and Windows. You’ll explore all the essentials of data science and linear algebra to perform data science tasks using packages such as SciPy, contrastive, scikit-learn, Rattle, and Rmixmod. Once you’re accustomed to all this, you’ll start with operations in data science such as cleaning, sorting, and data classification. You’ll move on to learning how to perform tasks such as clustering, regression, prediction, and building machine learning models and optimizing them. In addition to this, you’ll learn how to visualize data using the packages available for Julia, Python, and R.
Table of Contents (15 chapters)

A model's performance measure

In this chapter, we looked at several applications of linear models, including CAPM, the Fama-French 3-factor linear model, the Fama-French-Carhart 4-factor linear model, and the Fama-French 5-factor linear model. Obviously, CAPM is the simplest one since it only involves a market index as the explanatory variable. One question remains though: which model is the best? In other words, how do we rank these models and how is their performance measured? When running linear regressions, the output will show both the R2 and adjusted R2. When comparing models with different numbers of independent variables, the adjusted R2 is a better measure since it is adjusted by the number of input variables. However, note we should not depend only on the adjusted R2 since this is in the sample measure. In other words, a higher adjusted R2 simply means that based...