User Portlet
| Discussions |
|---|
| Fine! Thank you very much. Claude |
| As far as I know, *Classify* (and machine learning methods in general) are well-suited for **supervised classification**: the nature of several region is known, and they are used as training sets. So, using random drawn for that purpose is strange.... |
| I have to compare histogram of rational numbers, associated each time with 3 dataset and, sometimes, one of these datasets is empty. Example: consider Data = {{3/194, 1/194, 1/194, 1/194, 2/97, 3/194, 1/194, 2/97, 9/194, 1/97,... |
| Look at the eigenvalues of this symmetric matrix: Ch[x_] := CharacteristicPolynomial[{{qd + 2 qo + qv, -qo, -qo, -qd, 0, -qv}, {-qo, qd + 2 qo + qv, -qo, -qv, -qd, 0}, {-qo, -qo, qd + 2 qo + qv, 0, -qv, -qd}, {-qd,... |
| Hello, I extracted the fist 100 values of the original list, and the main problem is met again: the wavelet transfom with symlets is not invertible! More precisely, it depends on the number p of vanishing moments of SymletWavelet[p]: it... |
| I suppose each pair of numbers in data is: {time, number of death}? I considered two cases: (1) the sample consist in : numbers of death (2) the sample consist in : rate of death. and this sample obeys an Erlang distribution (FdrDir is the... |
| Yesss! The right manner is: Sol1 = {Apply[u, Flatten[{A, B}]], Apply[v, Flatten[{A, B}]]} /. solGlob; Thanks! |