Book Image

Data Science Projects with Python

By : Barbora stetinova
Book Image

Data Science Projects with Python

By: Barbora stetinova

Overview of this book

Data Science Projects with Python is designed to give you practical guidance on industry-standard data analysis and machine learning tools in Python, with the help of realistic data. The course will help you understand how you can use pandas and Matplotlib to critically examine a dataset with summary statistics and graphs and extract the insights you seek to derive. You will continue to build on your knowledge as you learn how to prepare data and feed it to machine learning algorithms, such as regularized logistic regression and random forest, using the scikit-learn package. You’ll discover how to tune the algorithms to provide the best predictions on new and, unseen data. As you delve into later chapters, you’ll be able to understand the working and output of these algorithms and gain insight into not only the predictive capabilities of the models but also their reasons for making these predictions. The codes for this course can be downloaded from https://github.com/TrainingByPackt/Data-Science-Projects-with-Python-eLearning.

Related Content you might be interested in

Current Title:

Small book image
Data Science Projects with Python
Barbora stetinova
Jun, 2019 | 6 hours 7 minutes
No titles found
Table of Contents (6 chapters)
Free Chapter
1
Data Exploration and Cleaning
Course Overview
Installation and Setup
Lesson Overview
Python and the Anaconda Package Management System
Different Types of Data Science Problems
Loading the Case Study Data with Jupyter and pandas
Getting Familiar with Data and Performing Data Cleaning
Boolean Masks
Data Quality Assurance and Exploration
Deep Dive: Categorical Features
Exploring the Financial History Features in the Dataset
Lesson Summary
2
Introduction to Scikit-Learn and Model Evaluation
Lesson Overview
Exploring the Response Variable and Concluding the Initial Exploration
Introduction to Scikit-Learn
Model Performance Metrics for Binary Classification
True Positive Rate, False Positive Rate, and Confusion Matrix
Obtaining Predicted Probabilities from a Trained Logistic Regression Model
Lesson Summary
3
Details of Logistic Regression and Feature Exploration
Lesson Overview
Examining the Relationships between Features and the Response
Finer Points of the F-test: Equivalence to t-test for Two Classes and Cautions
Univariate Feature Selection: What It Does and Doesn't Do
Generalized Linear Models (GLMs)
Lesson Summary
4
The Bias-Variance Trade-off
Lesson Overview
Estimating the Coefficients and Intercepts of Logistic Regression
Assumptions of Logistic Regression
How Many Features Should You Include?
Lasso (L1) and Ridge (L2) Regularization
Cross Validation: Choosing the Regularization Parameter and Other Hyperparameters
Reducing Overfitting on the Synthetic Data Classification Problem
Options for Logistic Regression in Scikit-Learn
Lesson Summary
5
Decision Trees and Random Forests
Lesson Overview
Decision Trees
Training Decision Trees: Node Impurity
Using Decision Trees: Advantages and Predicted Probabilities
Random Forests: Ensembles of Decision Trees
Fitting a Random Forest
Lesson Summary
6
Imputation of Missing Data, Financial Analysis, and Delivery to Client
Lesson Overview
Review of Modeling Results
Dealing with Missing Data: Imputation Strategies
Cleaning the Dataset
Mode and Random Imputation of PAY_1
A Predictive Model for PAY_1
Using the Imputation Model and Comparing it to Other Methods
Financial Analysis
Final Thoughts on Delivering the Predictive Model to the Client
Lesson Summary
Chapter 3
Details of Logistic Regression and Feature Exploration
Content Locked
Section 4
Univariate Feature Selection: What It Does and Doesn't Do
In the earlier videos, we have learned techniques for going through features one by one to see whether they have predictive power. This is a good first step, and if you already have features that are very predictive of the outcome variable, you may not need to spend much more time considering features before modeling. However, there are drawbacks to univariate feature selection. It does not consider the interactions between features. For example, what if the credit default rate is very high specifically for people with a certain education level and a certain range of credit limit? Here are the topics that we will cover now:
Previous Section
End of Section 4
Next Section