Applying Machine Learning

Applying Machine Learning

• start the course
• describe the two main types of error in machine learning models and the tradeoff between them
• describe how to use cross-validation to show how generalized a model is
• describe cross-validation in Python to obtain strong evaluation scores
• describe different metrics that can be used to evaluate binary classification models
• describe different metrics that can be used to evaluate non-binary classification models
• describe common evaluation metrics for evaluating classification models
• describe different metrics that can be used to evaluate regression models
• describe how to use Python to calculate common evaluation methods
• describe AWS machine learning
• set up an AWS environment and import data sources
• create a model with AWS
• set training criteria with AWS and train a model
• define bias, variance, and tradeoffs

Convolutional and Recurrent Neural Networks

Convolutional and Recurrent Neural Networks

• start the course
• describe convolutional neural networks, how they are different from regular neural networks, and how they are used
• describe the high level architecture of convolutional neural networks
• describe how convolution layers are set in convolutional neural networks
• describe how pooling layers work in convolutional neural networks
• describe some training considerations for convolutional neural networks and how training can differ from traditional neural networks
• describe regularization and how it applies to convolutional neural networks
• implement and train a convolutional neural network in TensorFlow
• perform regularizing to a convolutional neural network in TensorFlow
• describe recurrent neural networks, how they are different from regular neural networks, and how they are used
• describe the architecture of a recurrent neural network
• implement an LSTM network in TensorFlow
• use RNNs to perform time-series analysis in TensorFlow
• use TensorFlow to create a CNN that classifies images

Introduction to Machine Learning and Supervised Learning

Introduction to Machine Learning and Supervised Learning

• start the course
• define machine learning and how it can be used to solve a variety of problems
• define supervised machine learning
• describe the fundamentals of building machine learning models to solve a problem
• describe overfitting, how it can be a problem, and how to mitigate it
• evaluate machine learning models and compare them
• define the linear regression model for one and multiple variable problems
• describe the gradient descent algorithm for training linear regression models
• describe the k-nearest neighbor model and how to learn it
• describe decision tree models and how to learn decision trees using the C4.5 algorithm
• set up scikit-learn for Python
• import data, and perform basic tasks with scikit-learn for Python
• use scikit-learn to fit a linear regression model to a dataset
• use scikit-learn's k-nearest neighbor model
• use scikit-learn to fit a decision tree model to a dataset
• use scikit-learn and GraphViz to generate a decision tree model from a dataset
• use scikit-learn to calculate the precision and the recall of different machine learning models in Python
• fit a linear regression model to a dataset with scikit-learn and Python

Neural Networks

Neural Networks

• start the course
• describe neural networks and their capabilities
• describe how different neural networks are structured
• describe how cost functions are used to train neural networks
• describe activation functions and list different types of commonly used activation functions
• describe feedforward neural networks and the intuition behind calculating gradients in neural networks
• describe how to use backpropagation for more efficient neural network training
• describe batch learning and why it makes neural network training easier
• describe TensorFlow and its high-level architecture
• set up TensorFlow for use on a CPU
• import data into TensorFlow using built-in data sources and external data sources
• build and train a single-layer neural network in TensorFlow
• build and train a multilayer neural network in TensorFlow
• describe neural networks, network layers, cost functions, activation functions, and gradient descent

Supervised Learning Models

Supervised Learning Models

• start the course
• describe the difference between classification and regression models and the use for each of them
• describe how decision trees can be applied to regression problems
• describe the CART decision tree learning algorithm and how it's different from C4.5
• describe the random forests machine learning
• use scikit-learn to build a random forest model in Python
• describe the logistic regression model
• use scikit-learn to fit a logistic regression model
• describe support vector machine models
• describe how to use kernel methods with support vector machines to model more complex data
• use scikit-learn to train and support vector machines in Python
• describe the Naïve Bayes classifiers and how to train them
• use scikit-learn to fit a Naïve Bayes classifier in Python
• describe different supervised learning models in Python

Unsupervised Learning

Unsupervised Learning

• start the course
• describe unsupervised learning and some of the problems it can solve
• describe rule association and how the apriori algorithm performs this task
• use the apriori algorithm for rule association in Python
• describe clustering and the types of problems it applies to
• describe the k-means clustering algorithm
• use SciKit Learn to build clusters in python
• describe anomaly detection, the types of problems solved with anomaly detection, and some approaches to anomaly detection
• use scikit learn to perform anomaly detection
• describe the problems with dimensionality and why efforts to reduce dimensionality should be taken
• describe principal component analysis for dimensionality reduction
• use SciKit Learn to perform dimensionality reduction
• perform dimensionality reduction and clustering tasks in Python