Timeline for Anomaly detection
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 17, 2017 at 12:00 | comment | added | Rozita | I used EM algorithm instead. | |
| Oct 11, 2017 at 15:12 | comment | added | German Demidov | I am not saying that gamma distribution is the ideal choice. If I were you I would try to understand which stochastic process models your data and this understanding will tell you which distribution to use. I am not even sure that it is a correct way to deal with lm coefficients and are your linear models valid. But yes, if I wanted to fit this distribution with gamma - I would multiply it by -1. | |
| Oct 11, 2017 at 14:59 | comment | added | Rozita | One question here: Gamma distribution is for positive values and my lm coefficients are negative. Should I just simply multiplie them by -1 and then use gamma distribution? | |
| Oct 11, 2017 at 12:32 | comment | added | German Demidov | upd: your distribution (based on the plot) is not even close to normal. try to find a rationale to use smth like zero-inflated gamma distribution. | |
| Oct 11, 2017 at 11:54 | answer | added | RajeshS | timeline score: -1 | |
| Oct 11, 2017 at 9:29 | comment | added | German Demidov | At first, I would not rely on normality here. If you have 4266 + 229 + 160 points, there should be ~3100 points within the interval of 1SD under normality assumption, not >4000. I would find a probability model that "look alike" your distribution, fit it in a robust way and remove all results that have adjusted p-value < 0.05 (or any other threshold) | |
| Oct 11, 2017 at 9:26 | history | edited | Rozita | CC BY-SA 3.0 |
added 87 characters in body
|
| Oct 11, 2017 at 9:23 | comment | added | Rozita | How about using probablity function of normal distribution and applying an epsilon to decide which PF is good and which one is anomaly? | |
| Oct 11, 2017 at 9:20 | comment | added | Rozita | Yes it standard deviation. 2) it is skewed but so basically starts from the mean of the normal distribution. But what do you suggest to me to do? | |
| Oct 11, 2017 at 9:12 | comment | added | German Demidov | What is denoted as SD? If it is standard deviation, you either 1) estimated standard deviation in a non-robust way, or 2) your distribution of lm coefficient is not normal. | |
| Oct 11, 2017 at 8:57 | history | asked | Rozita | CC BY-SA 3.0 |