Please enable JavaScript.
Coggle requires JavaScript to display documents.
Machine Learning Algos (Reinforcement (Deep reenforcement learning (Deep Q…
Machine Learning Algos
Supervised learning
Algorithms
Classification / labeled
Binomial classification
Logistic regression
Support vector machine (SVM)
Multiclass classification
SVM
KNN
Euclidean
distance
Manhattan
distance
Random forest
Neural networks
Bayes
Regression
KNN
Anomaly
detection
Non parametric
Patametric
Decision trees
KNN
Steps
Collect Data
Row - Data item
Column - feature
pick features to maximize signal to noise ratio
Pick features those carry most information
Type
Integer (a number)
Binary (0/1 or Yes/No)
Training data
Test data
Train classifier
Python: ScikitLearn - Fit()
For image classifiers - Tensorflow for poets
Make predictions
Unsupervised learning
Steps
Algorithms
Clustering / un-labeled
K-Means
Hierarchical
Singular value decomposition (SVD)
Reinforcement
Exploration / exploitation
Epsilon greedy
Markov decision process (MDP)
Q-Learning
Policy learning
Policy function
Deep reenforcement learning
Deep Q networks (DQNs)
Asynchronous advantage actor-critic (A3C)
Neural network
Applicability
Pattern exists
No mathematical solution
Data available