Anyone working in "Machine Learning"?

[aussielong]

Actually you are wrong, weighted moving average is the first step in finding components of time series, if I remember correctly. Which is part of stats modelling, whose techniques are being incorporated into machine learning.

Completely wrong. You've got it arse about face again. Do you even know what a time series is?

Keep going dumbo. You'll probably trump your idiotic post about London robbers if you keep trying.

[Advocate]

Seems to me that Machine learning is bringing together the things I enjoy most workwise: algorithmic programming and statistical modelling.
Especially with Bayesian techniques coming to the fore recently, which are more computationally intensive than frequentist methods.
Anyone working in this area commercially?

This was on radio 4 this morning! Yes, I listen to radio 4, aren't I posh?

Sent from my tweeting foot massager.

[sasguru]

Completely wrong. You've got it arse about face again. Do you even know what a time series is?

.

I've done a course on time series, yes.
Do give us the benefit of your extensive stats knowledge, will you?
So I can rip you to shreds.
meanwhile I have to go and do some actual stats work.

[formant]

I sorta do. In Back-End Speech Recognition. I'm not a developer though. I do the linguistic and phonetic fine tuning.

[sasguru]

I can hear the sound of frantic googling, getting increasingly desperate as Oz the idiot realises that weighted moving averages are instrumental in getting the trend component in time series.

Typical Ozzie: all mouth, no trousers.

[aussielong]

I've done a course on time series, yes.
Do give us the benefit of your extensive stats knowledge, will you?
So I can rip you to shreds.
meanwhile I have to go and do some actual stats work.

Oh you did a course on it did you grandad?



You do a lot of courses don't you...

[sasguru]

Oh you did a course on it did you grandad?



You do a lot of courses don't you...

Conceded defeat already? What a plonker.
Time series was covered on my MSc more than 10 years ago, its not my speciality nowadays.
I have a suggestion: get a decent qualification and some experience, then you can move on from your $30an hour VB scripting interspersed with $10 an hour bar-tending.

[aussielong]

Conceded defeat already? What a plonker.
Time series was covered on my MSc more than 10 years ago, its not my speciality nowadays.
I have a suggestion: get a decent qualification and some experience, then you can move on from your $30an hour VB scripting interspersed with $10 an hour bar-tending.

How do I know if your MSc is a good one? It might not be worth the paper it was written on for all I know.

The evidence would suggest that. You're a one trick pony, what.

[sasguru]

How do I know if your MSc is a good one? It might not be worth the paper it was written on for all I know.

The evidence would suggest that. You're a one trick pony, what.

MSc Stats, Sheffield University.
It's true you are a 2-trick pony: VB scripting and bar-tending. That I concede.

[DimPrawn]

I've done a course on time series, yes.
Do give us the benefit of your extensive stats knowledge, will you?
So I can rip you to shreds.
meanwhile I have to go and do some actual stats work.

Q1
(in your own words)

Compare and contrast correlation and cointegration with respect to two timeseries, and how this relates to fractal dimensionality.