Linear models diffuse in many different scientific and business applications and can be found, under different functions, in quite a number of different Python packages. We have selected a few for use in this book. Among them, Statsmodels is our choice for illustrating the statistical properties of models, and Scikit-learn is instead the package we recommend for easily and seamlessly preparing data, building models, and deploying them. We will present models built with Statsmodels exclusively to illustrate the statistical properties of the linear models, resorting to Scikit-learn to demonstrate how to approach modeling from a data science point of view.
NumPy, which is Travis Oliphant's creation, is at the core of every analytical solution in the Python language. It provides the user with multidimensional arrays, along with a large set of functions to operate multiple mathematical operations on these arrays. Arrays are blocks of data arranged along multiple dimensions and that implement mathematical vectors and matrices. Arrays are useful not just for storing data, but also for fast matrix operations (vectorization), which are indispensable when you wish to solve ad hoc data science problems.
In the book, we are primarily going to use the module linalg from NumPy; being a collection of linear algebra functions, it will provide help in explaining the nuts and bolts of the algorithm:
Website: http://www.numpy.org/
Import conventions:
import numpy as npVersion at the time of print:
1.9.2Suggested install command:
pip install numpy
An original project by Travis Oliphant, Pearu Peterson, and Eric Jones, SciPy completes NumPy's functionalities, offering a larger variety of scientific algorithms for linear algebra, sparse matrices, signal and image processing, optimization, fast Fourier transformation, and much more.
The scipy.optimize package provides several commonly used optimization algorithms, used to detail how a linear model can be estimated using different optimization approaches:
Website: http://www.scipy.org/
Import conventions:
import scipy as spVersion at time of print:
0.16.0Suggested install command:
pip install scipy
Previously part of Scikit, Statsmodels has been thought to be a complement to SciPy statistical functions. It features generalized linear models, discrete choice models, time series analysis, and a series of descriptive statistics as well as parametric and nonparametric tests.
In Statsmodels, we will use the statsmodels.api and statsmodels.formula.api modules, which provide functions for fitting linear models by providing both input matrices and formula's specifications:
Website: http:/statsmodels.sourceforge.net/
Import conventions:
import statsmodels.api as smandimport statsmodels.formula.api as smfVersion at the time of print:
0.6.1Suggested install command:
pip install statsmodels
Started as part of the SciPy Toolkits (SciKits), Scikit-learn is the core of data science operations on Python. It offers all that you may need in terms of data preprocessing, supervised and unsupervised learning, model selection, validation, and error metrics. Expect us to talk at length about this package throughout the book.
Scikit-learn started in 2007 as a Google Summer of Code project by David Cournapeau. Since 2013, it has been taken over by the researchers at INRA (French Institute for Research in Computer Science and Automation).
Scikit-learn offers modules for data processing (sklearn.preprocessing, sklearn.feature_extraction), model selection, and validation (sklearn.cross_validation, sklearn.grid_search, and sklearn.metrics) and a complete set of methods (sklearn.linear_model) in which the target value, being both a number or a probability, is expected to be a linear combination of the input variables:
Website: http://scikit-learn.org/stable/
Import conventions: None; modules are usually imported separately
Version at the time of print:
0.16.1Suggested install command:
pip install scikit-learn