Description:This 6-video course focuses on understanding Google's TensorFlow estimators, and showing learners how they simplify the task of building simple linear and logistic regression models for machine learning solutions. As a prerequisite, learners should have a basic understanding of ML (machine learning), and basic experience programming in Python. Though not required, familiarity with the Scikit-learn library and the Keras API will simplify the labs part of this course. First, you will learn how TensorFlow estimators abstract many of the details in creating a neural network, and you will then learn that you no longer need to define the type of neural network model, nor will you need to add definitions to layer. When using an estimator, learners only need to feed in training and validation data. In the course labs, you will build both a linear regression model and a classifier by using TensorFlow estimators. Finally, you will learn how to evaluate your model using the prebuilt methods available in the estimator.
Target Audience:
Duration: 00:36
Description:
Logistic regression is a technique used to estimate the probability of an outcome for machine learning solutions. In this 10-video course, learners discover the concepts and explore how logistic regression is used to predict categorical outcomes. Key concepts covered here include the qualities of a logistic regression S-curve and the kind of data it can model; learning how a logistic regression can be used to perform classification tasks; and how to compare logistic regression with linear regression. Next, you will learn how neural networks can be used to perform a logistic regression; how to prepare a data set to build, train, and evaluate a logistic regression model in Scikit Learn; and how to use a logistic regression model to perform a classification task and evaluate the performance of the model. Learners observe how to prepare a data set to build, train, and evaluate a Keras sequential model, and how to build, train, and validate Keras models by defining various components, including activation functions, optimizers and the loss function.
Target Audience:
Duration: 00:58
Description:
Several factors usually influence an outcome, and users need to consider all of those by using regression. Explore machine learning techniques and risks involved when using multiple factors for linear regression in this 11-video course. Key concepts covered here include reasons to use multiple features in a regression, and techniques involved in creating a model; preparing a data set containing multiple features for training and evaluating a linear regression model; and how to configure, train, and evaluate the linear regression model. Next, learn to create a data set with multiple features in a form which can be fed to a neural network for training and validation; learn the architecture for a Keras sequential model and setting training parameters; and learn to make predictions on test data and examine the metrics to gauge the quality of the neural network model. Learn to use Pandas and Seaborn to view correlations and enumerate risks; apply the principle of parsimonious regression to rebuild the linear regression model; and build a Keras model after selecting only important features from a data set.
Target Audience:
Duration: 01:11
Description:
Learn how to use the Scikit Learn and Keras libraries to build a linear regression model to predict a house's price in this 8-video course, and learn steps involved in preparing data and configuring regression models. Key concepts covered here include using the Pandas library to load a data set in the form of a CSV file for consumption by a linear regression model; creating training and validation sets for a regression model; and how to configure a linear regression model and train and validate it, view the metrics for the model, and visualize it by using Matplotlib. Next, learn to install the Keras library and prepare the data set for consumption by a Keras model; learn the architecture for a Keras sequential model and initialize it; and compile a Keras sequential model by defining loss function and optimizer and train it to get optimal values for weights and biases. Finally, evaluate a Keras sequential model by using it to make predictions on test data; and work with training sets and the Keras sequential model for machine learning solutions.
Target Audience:
Duration: 00:42
Description:
Machine learning (ML) is everywhere these days, often invisible to most of us. In this 12-video course, you will discover one of the fundamental problems in the world of ML: linear regression. Explore how this is solved with classic ML as well as neural networks. Key concepts covered here include how regression can be used to represent a relationship between two variables; applications of regression, and why it is used to make predictions; and how to evaluate the quality of a regression model by measuring its loss. Next, learn techniques used to make predictions with regression models; compare classic ML and deep learning techniques to perform a regression; and observe various components of a neural network, such as neurons and layers and how they fit together. You will learn the two types of functions used in a neuron and their individual roles; steps involved in calculating the optimal weights and biases of a neural network; and the technique of gradient descent optimization, needed to find optimal parameters for a neural network.
Target Audience:
Duration: 01:19