Course Code: geamlm
Duration: 28 hours
Prerequisites:

There are no specific requirements needed to attend this course.

Course Outline:

Part I

Processing and analysis of geo-spatial data

Introduction to geo-spatial data

  • R and GIS
  • Types of spatial data
  • Storage and display
  • Spatial Data analysis

Classes for geo-spatial data in R

  • Overview of R packages to handle Spatial data
  • Classes and Methods
  • Spatial points
  • Raster objects

Import/Export geo-spatial data to/from R

  • Overview of R packages to import data to R
  • Coordinate reference systems
  • Vector file formats
  • Raster file formats
  • Other formats
  • Import/Export interfaces to open source data

Manipulation on geo-spatial data

  • Calculating distance
  • Triangulation, Interpolation
  • Calculating area
  • Calculating coverage

Visualizations of geo-spatial data

  • Overview of R packages for Visualization geo-spatial data
  • Basic mapping with ggplot2 and ggmap
  • Point and polygon data
  • Raster data and color
  • Dynamic mapping 

Part II

Machine learning and Time series Forecasting

Introduction to Time series Forecasting

  • Time series characteristics – Trends, Seasonality
  • Stationarity and stochastic trends
  • Measuring forecast accuracy
  • Stability of forecast accuracy
  • Simple methods of time series forecasting

Introduction to Machine learning

  • Applications of machine learning
  • Supervised and unsupervised learning
  • Bias vs Variance dilemma and Overfitting
  • Measures of Accuracy, Cross-Validation
  • Increasing predictive ability – model tuning
  • Ensemble learning

Traditional methods of time series forecasting

  • Decomposition of time series – X13 Arima Seats
  • Exponential Smoothing in State Space
  • ARIMA and seasonal ARIMA
  • Error correction models
  • Structural models for time series forecasting
  • VAR models
  • Forecasting hierarchical time series data

Machine learning algorithms

  • Dimensionality reduction
  • Regularization of linear regression
  • Decision trees – classification problems
  • Extreme Gradient boosting
  • Support Vector Machines
  • Neural networks
  • Deep networks
  • Clustering