Machine Learning

Currently this is kind of a stub article outlining some of the areas to be covered.

• Machine Learning Reference 1
• Machine Learning Reference 2
• Machine Learning Reference 3
• Machine Learning Reference 4 <== just starting
• Machine Learning Reference 5
• Machine Learning Reference 6

“signal” or “feedback” based

• Supervised learning
  • inputs and outputs
  • learn general rule to map them
  • examples:
    • Image Classification - train with images and labels
    • Market Prediction/Regression - train with
• unsupervised learning
  • no labels
  • needs to find structure
  • discovering hidden patterns

  • clustering population in groups

  • Clustering - separate data into clusters
  • High Dimension Visualization - visualize high dimension data

  • Generative Models - captures probability distribution

    ----- supervised / unsupervised image ------
    

  • Semi-supervised learning
    • lot of input data, some of it is labeled
  • Reinforcement learning
    • rewards and punishments as feedback

Desired Output Based

• Classification ….
  • multi-label classification …
  • usually supervised …
• Regression
  • supervised
  • outputs are continuous rather than discrete

  —- classification / regression image ———

• Clustering
  • inputs divided into groups
  • usually unsupervised
  • groups not known ahead of time
• Density estimation - distribution of inputs in some space
• Dimensionality reduction
  • map inputs into lower dimensional space to simplify
  • topic modeling
• Model ( hypothesis ) - representation learned by applying ML algorithms to data
• feature - measurable property of data
• feature vector - set of features, input to a model

• feature extractor - extracts relevant features from data

• Target (Label) - the value you want to predict

• training
  • building a model
  • given outputs(labels) and inputs inputs(features)
• prediction
  • using a model
  • inputs are used to find outputs

—- training / prediction image —-

• computational learning theory

AI research three approaches

• computational psychology
• computational philosophy
• computer science

Types of Agents

• Simple Reflex Agents
• Model-Based Reflex Agents
• Goal-Based Agents
• Utility-Based Agents
• Learning Agent

Example of Supervised Learning Algorithms:

• Linear Regression
• Nearest Neighbor
• Guassian Naive Bayes
• Decision Trees
• Support Vector Machine (SVM)
• Random Forest

Data for unsupervised learning:

• “Unstructured data: May contain noisy(meaningless) data, missing values or unknown data”
• Unlabeled data ( not label / output )

Types of Unsupervised Learning :

• Clustering
• Association

Some unsupervised learning algorithms:

• K-Means Clustering
• DBSCAN – Density-Based Spatial Clustering of Applications with Noise
• BIRCH – Balanced Iterative Reducing and Clustering using Hierarchies
• Hierarchical Clustering

Reinforcement Learning:

• Some algorithms:
  • Temporal Difference (TD)
  • Q-Learning
  • Deep Adversarial Networks
• decisions are dependent

Supervised learning algorithm types:

• Classification - puts it in a group
• Regression - assigns a quantity/value Unsupervised learning algorithm types :
• Clustering
• Association