Modelling the NASDAQ100

• 

  Data

• 

  log-transformed data

• 

  discretized data

Inspiration

Trying to figure out what Markov chains are useful for

What it does

Ideally: in RStudio, from a discrete-time continuous-state data set, first discretize the data, build a markov chain, fit the model and conduct testing to see if it's appropriate. In practice: I only managed to discretize the data

How we built it

From historical data in CSV -> loaded into RStudio -> Rcode

Challenges we ran into

--- doing everything from scratch, no previous knowledge of how to simulate or translate theory to code

Checking assumptions (normalized increments)

• log-transform
• discretizing with fixed mesh size -- incorrect mesh size, incorrect discretizing procedure

Accomplishments that we're proud of

Accomplishing the first step Keep trying Making friends Having a good time at HackBrunel

What we learned

doing new things takes time, usually a lot more time than you thought having friends to bounce ideas off of really gets you inspired

What's next for Modelling the NASDAQ100

All the other steps: Build markov chain, fit the model, test if it's actually appropriate

Built With

• r