Please enable JavaScript.
Coggle requires JavaScript to display documents.
Machine learning: learning properties about data and apply it to new data
Machine learning: learning properties about data and apply it to new data
Supervised learning: inputs and targets (labels)
Classification
labels are discrete values, e.g. handwritten digits
Regression
labels are continuous values, e.g. the length of salmon fish
Unsupervised learning
Clustering
Discover groups of similar attributes within data
Density estimation
determine distribution of data within input space
Scikit-Learn estimators, one layer perceptrons:
http://scikit-learn.org/stable/tutorial/machine_learning_map/index.html
support vector machines
Gradient descent
Random forest
nearest neighbors
The difference between classification and regression is the activation function
http://scikit-learn.org/stable/tutorial/statistical_inference/supervised_learning.html
Neural networks
Can do all that an estimator can do
Is different from logistic regression, by having more than one layer
Neural network is a multilayer perceptron
http://scikit-learn.org/stable/modules/neural_networks_supervised.html
TensorFlow
Optimizers
gradient descent
momentum optimizer
And many more
Neural networks
Scikit-learn neural nets, based of multi-layer perceptrons:
http://scikit-learn.org/stable/modules/neural_networks_supervised.html
regression: MLPRegressor
classification: MLPClassifier
Statistical learning, rule-based learning, learning classifier system
Varies the parameters to minimize the cost
Theano