Timeline for What statistical test can I use to detect clumping?
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 25, 2011 at 4:53 | vote | accept | Mike Furlender | ||
| May 24, 2011 at 2:28 | comment | added | bill_080 | @Mike: I updated the graph (Edit 2) to show more "windows" for comparison. The same ideas apply to hourly data as they do to daily data. | |
| May 24, 2011 at 2:24 | history | edited | bill_080 | CC BY-SA 3.0 |
added 1513 characters in body
|
| May 23, 2011 at 22:23 | comment | added | Mike Furlender | @Bill I guess what I should do is to make a "cutoff" within, say, hourly windows, beyond which the result is unacceptable. | |
| May 23, 2011 at 22:20 | comment | added | Mike Furlender | @Bill You know what, maybe you're right and I'm over complicating it. I gotta think about it. I think perhaps what I was trying to solve was "clusters of clusters" instead of just clusters. Hm.. | |
| May 23, 2011 at 22:15 | comment | added | Mike Furlender | @Bill The problem is that if I make the "window" 1 day, then I get a number that represents the amount of trades that took place in that day. This is not what I am referring to as the "cluster number." I need to know the distribution of the variance within that day. For instance if I had 200 trades in 1 day, 100 of which happened at the very beginning and 100 of which happened at the very end, then with your method I would get "200" as the "clustering number." But this tells me nothing of their distribution internally. | |
| May 23, 2011 at 22:05 | vote | accept | Mike Furlender | ||
| May 23, 2011 at 22:06 | |||||
| May 23, 2011 at 20:37 | comment | added | bill_080 | @Mike: You need to concentrate the idea of "the window". In the above case, "the window" is 1 day. And for trading purposes, you would react each day to the previous day. The full graph above is 180 days. However, if you want to use 180 days as "the window" then the total number of trades in the previous 180 days would be your "cluster number" (in this case, there were a total of 404 trades in those 180 days, giving you a cluster number of 404). This 180-day window would advance each day, dropping off the trades that are 181 days ago and picking up the latest trades. | |
| May 23, 2011 at 20:20 | comment | added | Mike Furlender | @bill I need to create 1 number for the whole graph because I intend to automatically react to this number by entering or not entering another (unrelated) trade. I can take the average of every clustering number within the graph and I guess that would do it, but it seems like an imperfect method. I may have to take the standard deviation between these clustering numbers.... perhaps that is the way. | |
| May 23, 2011 at 20:08 | comment | added | bill_080 | @Mike: Each day has a "cluster number". If you need one number for the whole graph, then the "rolling window" is the whole graph. | |
| May 23, 2011 at 20:07 | history | edited | bill_080 | CC BY-SA 3.0 |
edited body
|
| May 23, 2011 at 20:06 | comment | added | bill_080 | @Mike: The "clustering number" would simply be the number of trades per day. I updated the above graph. "Low clustering" might be a value of 1 trade or less per day (the green line or below). "High clustering" might be a value of 10 or more trades per day (the red line or above). "Normal clustering" might be anything in between the red and green lines. You can have as many "degrees of clustering" (horizontal lines) as you like. So, you can spot any degree of "clustering" simply by looking at the rolling "trades per day" number. | |
| May 23, 2011 at 19:44 | comment | added | Mike Furlender | @bill, Good idea- but how do I then assign 1 number to the whole graph you posted? (I need to do this). | |
| May 23, 2011 at 19:35 | history | answered | bill_080 | CC BY-SA 3.0 |