Instance Based Learning

In this lecture series professor explains with instance based learning It is nothing but learning based on instances of data *Lazy Learning- Learn at last after creating a model - Knn Neighbour *Easy Learning - Learn as soon as we receive the data


  • Started with storing things in DB and querying but what will happen when no such combinations are present
  • Find the nearest closest value/similarity(distance)
  • Use distance formula to calculate the nearest distances
    • Euclidian
    • Manhattan are two such formulas for calculation of distances.
  • If there are more nearby values then calculate the averages
  • Curse of Dimensionality- More data more dimensions
  • Locally Weighted Regression