Recently Written
Categories
(options)
- R cheat page
- Most functions are kept in utils.R
- LReg.R Does Local Regression smoothing used within getHits …
- getHits.R
- Very useful for locating the Hit times in a segment.
- LReg1.R
- For fine tunning location of Hit times.
- plotTs.R
- Use it in combination with getHits and LReg1. Here is an example.
- The SVM
“Normal” & “Random” .wav files for the SVM, also all associated .txt files.
- Naked amplitudes are not very good. Notice that even the data used for training the svm is not perfectly classified!. MFCCs are needed. Here is the R code
New Utilities with maxnkeys=60 NewUtils.txt
- First Draft MFCC SVM
“Normal” & “Random” .wav files for the MFCC SVM, also all associated .txt files.
mfccSVM.R GETS 70% accuracy!! (notice the change of pars for MFCC)
11,4,2009
- 100% Accuracy Discrimination Between A Key and L Key.
AFC+MFCC.R Individual Parameter Testing , SEARCHAFC+MFC.R runif Parameter Testing
Parameters.txt 100% Accuracy!!!
SVMClassification.jpeg SVM Classification with 28/100 support vectors.
11,18,2009
- 100% Accuracy Discrimination Between Q Key and P Key.
PQAFC+MFCC.R Individual Parameter Testing , QPSEARCHAFC+MFC.R runif Parameter Testing
HITS FOR Q and P.txt 100% Accuracy!!!
QPSVMClassification.jpeg SVM Classification with 43/100 support vectors.
3,3,2010
- 82.22% Accuracy Discrimination Between ALQP Keys Using Binary Classificatoin.
_________________________________________
« Papers | KeysTrail | Blog »


Recent Comments