Building Predictive Models from Scratch Using Python
About This Course
The course is an introduction to classical Machine Learning technique using Python. It introduces learners to basic machine learning steps from data preparation to evaluation of machine learning models. Learners will learn and build two classical machine learning models namely Linear Regression and Logistic Regression for continuous and categorical data respectively. Learners will learn how to process data using Pandas library in Python as well as to visualize those data using Seaborn and Matplotlib. On top of that, they will write the functions to build machine learning models using Numpy. Instead of using Scikit-Learn Library, learners will write their own machine learning functions to gain deeper understanding how such library functions. At the end, they will learn some metrics to evaluate their machine learning models.
What You'll Learn
* Create a Panda’s DataFrame and selecting data from DataFrame
* Use library to read Comma-separated values (CSV) or EXCEL file
* Split data randomly into training set and testing set
* Normalize data using min-max normalization
* Give example of linear regression and classification
* Write objective function of linear regression
* Implement Gradient Descent algorithm for optimisation
* Train linear regression model using gradient descent
* Transform data for higher order features
* Evaluate linear regression model using R-squared and mean-squared-error
* Evaluate and choose learning rate
* Plot cost function over iteration time
* Plot linear regression
* Write objective function of logistic regression
* Use logistic regression to calculate probabilities of binary classification
* Train logistic regression model
* Split data into training, validation, and testing set
* Visualize non-linear decision boundary
* Classify multi-class problems using one-vs-all technique
* Calculate confusion matrix, precision, and recall
Entry Requirements
Participants should possess a basic understanding of the Python programming language and should have gone through the Fundamentals in Python (Basic) course and Part 1 of this course.