arrow_back Distributed Computing Abstractions for Big Data Science
Four horsemen of the IoT arrow_forward
Data Simulation as a means to intuitively grasp Statistics and it's direct application to prediction problems
Submitted by Lakshman Prasad (@becomingguru) on Tuesday, 7 June 2016
Section: Full talk Technical level: Beginner
Whenever there is data, there is meta-data about the data itself characterised in the form of Statistics.
Contemporary statistics charecterizes these using mathematical distributions such as the p curve, t curve. To fully comprehend and/or derive these takes quite an analytical prowress. Even then, their applicability holds on the validity of the inherent assumptions which may not hold in a complex real world situation.
Since computers can simply generate data as specified in a model or in random, there’s an alternative valid way to look through and intutively grasp some statistics based on simulation.
This session intends to enable intutive grasping of some fundamentals of statistics and further it’s direct application in some complex real world situations where analytical approaches might prove too hard.
- Fundamentals of Statistics
- Normal Distribution
- The p curve
- Negative Binomial distribution
- Analytical approach to solving “number of heads in n tosses”
- Simulation approach to solve the same
- Randomn number generators
- Perception of random vs being random
- Casino fallacy
- Other simulation approaches for statistical problems
- Complex Real World problem
- An attempt at analytical solution
- Corresponding simulation solution
Lakshman Prasad works as a software architect at a management consulting firm and has been working on data and software for the better part of the decade.