Search results

I did not see any courses specifically targeting financial machine learning and forecasting. I did see some courses on things like packaging optimization, which are similar to some previous Kaggle competitions. But not specifically geared towards Quantitative finance or trading. If you have...

Interesting. Thanks for sharing. I think what they are bringing to the table is simply more raw power and complexity. What's of significantly more value, is the knowledge of how to apply these tools successfully (or not). I don't really think they are selling the tools based on financial...

People often refer to ML systems as opaque black boxes. I’d be curious to see a successful ML application whereby one could not possibly understand why it is successful, or how it is processing the data, given all of the information it utilizes as well as the underlying architecture. Any examples?

That's pretty interesting. Most published and common sense approaches I've seen keep the higher quantile and throw out the rest. That tries to mitigate some of the over-fitting by averaging over several series, rather than pick one best. The initial intuition about picking the middle set, would...

The 1st and 2nd course. Would have completed the others, but there were problems with the grader (and lack of class support) and I only had a week or two to audit all. Decided to go back in the future and complete when they cleaned up the outstanding problems. Despite the issues, I thought they...

Did you finish the course(s)? What are your thoughts so far? I thought they were quite good, except for the grader issues. Finished early without feedback, so not sure if they fixed them.

If this is for your master's thesis, I recommend you to familiarize yourself with the field of optimization. Particularly, search based methods would be useful here. There are also plenty of existing published papers out there to research.

Where did you get that Pi was based on Martin Armstong? I loved Pi, but would be very surprised to find out this was true. Far as I know, it was a work of fiction. As an aside, that out of sample point I posted a while back looks like it already failed. And that's just one OOS point already...

That really depends on how you define what 'shitty' or 'good' is. Notice you never did define it. 'Looking' at a graph characteristic in segments, is not the same as using a single quantitative metric to determine the outcome(s). In my case I simply used terminal wealth as a proxy (which isn't...

It's very easy to show that both classrooms will not necessarily end up with the same selection of models. All you need to do is imagine that model_n did very poorly on the first segment, and did fantastically well on the second segment. So well, in fact, that the model_n performance on the...

Walk forward and segementing folds of data is immensly useful if you try to understand the modelling process from a complexity/generalization perspective. Just imagine that you have a very finely optimized model on data segment 1, that performs spectacularly-- then (the same fitted model) on...

This contains a good applied illustration using portfolio weightings, reported performance, and the potential to mislead observers, depending on how data is aggregated.

Understandable. If there is a public data source (even a chart provider) to compare to , it can give you some sense of validation --One reason I thought divide and conquer might at least mitigate the problems (maybe so, maybe not) . If you find any good historical providers with ETFs (paid or...

FWIW, I can still easily download a few thousand EOD symbols from Yahoo finance no problem. I have a set of about 1900 ETFs (csv/zoo format). But from casual inspection, there are lots of gaps and outliers everywhere. I think it is cleanable and can imagine a simple script to adjust for...

for R quantmod users, I have not verified any data integrity issues yet (as mentioned by others), but joshuaulrich has posted a working solution and how to upload at github . Date range worked fine (e.g. SPY back to 1993). Visually looks fine.

Using R, even if you copy the exact link with crumb, e.g. https://query1.finance.yahoo.com/v7/finance/download/MSFT?period1=1463461200&period2=1494910800&interval=1d&events=history&crumb=XBOjLoYE5PS or, at R cmd line, SPY <-...

Just to play the Devil's advocate, one could have projected the green line pre-08, and extrapolated the bullish projection forward, using the same logic. The gain did continue -- with a slight hiccup along the way...

As was previously mentioned (GAT) python, matlab, and R are all great alternatives to Excel for this type of analysis. If you can learn and utilize any of them, you'll seldom find reason to go back to Excel for modelling. 1) Many ways to do this. Pretty much the goal of statistical modelling...

That's counter-intuitive, but I could see it making sense if you are simply averaging many draws of individual stock picks as opposed to holding an index. I tend to look at it that the individual draws, would have much wider (expected volatility and) confidence intervals over the future, than a...

@wintergasp I like that response. Could you elaborate a little more? @djames There are ways to combine by importance weighting, but it might be better to start by clearly understanding your objective and constraints and building an objective function from there (like simples pointed out)...