Supervised Learning with Python

Gain a thorough understanding of supervised learning algorithms by developing use cases with Python. You will study supervised learning concepts, Python code, datasets, best practices, resolution of common issues and pitfalls, and practical knowledge of implementing algorithms for structured as well as text and images datasets.

You’ll start with an introduction to machine learning, highlighting the differences between supervised, semi-supervised and unsupervised learning. In the following chapters you’ll study regression and classification problems, mathematics behind them, algorithms like Linear Regression, Logistic Regression, Decision Tree, KNN, Naïve Bayes, and advanced algorithms like Random Forest, SVM, Gradient Boosting and Neural Networks. Python implementation is provided for all the algorithms. You’ll conclude with an end-to-end model development process including deployment and maintenance of the model.After reading Supervised Learning with Python you’ll have a broad understanding of supervised learning and its practical implementation, and be able to run the code and extend it in an innovative manner.

WHAT YOU'LL LEARN

* Review the fundamental building blocks and concepts of supervised learning using Python
* Develop supervised learning solutions for structured data as well as text and images
* Solve issues around overfitting, feature engineering, data cleansing, and cross-validation for building best fit models
* Understand the end-to-end model cycle from business problem definition to model deployment and model maintenance
* Avoid the common pitfalls and adhere to best practices while creating a supervised learning model using Python

WHO THIS BOOK IS FOR

Data scientists or data analysts interested in best practices and standards for supervised learning, and using classification algorithms and regression techniques to develop predictive models.VAIBHAV VERDHAN has 12+ years of experience in Data Science, Machine Learning and Artificial Intelligence. An MBA with engineering background, he is a hands-on technical expert with acumen to assimilate and analyse data. He has led multiple engagements in ML and AI across geographies and across retail, telecom, manufacturing, energy and utilities domains. Currently he resides in Ireland with his family and is working as a Principal Data Scientist. Chapter 1: Introduction to Supervised Learning

Chapter Goal: Start the journey of the readers on supervised learning

No of pages: 30-40

Sub -Topics

1. Machine learning and how is it different from software engineering?

2. Discuss reasons for machine learning being popular

3. Compare between supervised, semi-supervised and unsupervised algorithms

4. Statistical methods to get significant variables

5. The use cases of machine learning and respective use cases for each of supervised, semi-supervised and unsupervised algorithms

Chapter 2: Supervised Learning for Regression Analysis

Chapter Goal: Embrace the core concepts of supervised learning to predict continuous variables

No of pages: 40-50

Sub - Topics

1. Supervised learning algorithms for predicting continuous variables

2. Explain mathematics behind the algorithms

3. Develop Python solution using linear regression, decision tree, random forest, SVM and neural network

4. Measure the performance of the algorithms using r square, RMSE etc.

5. Compare and contrast the performance of all the algorithms

6. Discuss the best practices and the common issues faced like data cleaning, null values etc.

Chapter 3: Supervised Learning for Classification Problems

Chapter Goal: Discuss the concepts of supervised learning for solving classification problems

No of pages : 30-40

Sub - Topics:

1. Discuss classification problems for supervised learning

2. Examine logistic regression, decision tree, random forest, knn and naïve Bayes. Understand the statistics and mathematics behind each

3. Discuss ROC curve, akike value, confusion matrix, precision/recall etc

4. Compare the performance of all the algorithms

5. Discuss the tips and tricks, best practices and common pitfalls like a bias-variance tradeoff, data imbalance etc.

Chapter 4: Supervised Learning for Classification Problems-Advanced

Chapter Goal: cover advanced classification algorithms for supervised learning algorithms

No of pages:30-40

Sub - Topics:

1. Refresh classification problems for supervised learning

2. Examine gradient boosting and extreme gradient boosting, support vector machine and neural network

3. Compare the performance of all the algorithms

4. Discuss the best practices and common pitfalls, tips and tricks

Chapter 5: End-to-End Model Deployment

Chapter Goal: guide the reader on the end-to-end process of deploying a supervised learning model in production

No of pages:25-30

1. Meaning of model deployment

2. Various steps in the model deployment process

3. Preparations to be made like settings, environment etc.

4. Various use cases in the deployment

5. Practical tips in model deployment