Skip to main content

Table 2 The sensitivity of the combinations of various classifiers and feature reduction algorithms.

From: Context based mixture model for cell phase identification in automated fluorescence microscopy

classifier

FR algorithm

Sensitivity (confidence Interval) (at 90% confidence level)

  

class 1

Class 2

class 3

class 4

CBMM

PCA

0.9390 (0.9100,0.9680)

0.7602 (0.6786,0.8419)

0.9526 (0.9196,0.9856)

0.8575 (0.8026,0.9125)

 

LDA

0.9220 (0.8966,0.9773)

0.7567 (0.8096,0.9237)

0.9489 (0.9462,1.000)

0.9460 (0.9101,0.9818)

 

MMC

0.9202 (0.8979,0.9425)

0.7733 (0.6724,0.8742)

0.9289 (0.9086,0.9491)

0.8445 (0.7774,0.9116)

 

SDAFS

0.9203 (0.8733,0.9401)

0.5537 (0.3536,0.4756)

0.9588 (0.9367,0.9817)

0.8786 (0.8191,0.9381)

 

MIFS

0.9067 (0.8733,0.9401)

0.4146 (0.3536,0.4756)

0.9592 (0.9367,0.9817)

0.8786 (0.8191,0.9381)

 

GA

0.9234 (0.8869,0.9599)

0.5812 (0.4748,0.6877)

0.9611 (0.9344,0.9878)

0.8743 (0.8273,0.9213)

SVM

PCA

0.8790 (0.8542,0.9038)

0.8025 (0.7328,0.8722)

0.9192 (0.9067,0.9317)

0.8744 (0.8441,0.9047)

 

LDA

0.6254 (0.6128,0.6380)

NaN

0.8411 (0.8185,0.8636)

NaN

 

MMC

0.8658 (0.8376,0.8941)

0.7694 (0.7061,0.8238)

0.8919 (0.8608,0.9230)

0.8221 (0.7882,0.8560)

 

SDAFS

0.8731 (0.8519,0.8943)

0.7767 (0.7043,0.8491)

0.9267 (0.9086,0.9447)

0.8966 (0.8734,0.9200)

 

MIFS

0.8819 (0.8607,0.9030)

0.7814 (0.7113,0.8515)

0.9421 (0.9247,0.9594)

0.8899 (0.8640,0.9157)

 

GA

0.8936 (0.8727,0.9146)

0.8332 (0.7689,0.8975)

0.9439 (0.9263,0.9615)

0.8852 (0.8480,0.9226)

KNN

PCA

0.8570 (0.8328,0.8811)

0.8880 (0.8273,0.9488)

0.9365 (0.9215,0.9515)

0.8753 (0.8467,0.9041)

 

LDA

0.4596 (0.4455,0.4736)

0.1751 (0.1406,0.2096)

0.4189 (0.3650,0.4729)

03689 (0.2580,0.4800)

 

MMC

0.8590 (0.8367,0.8812)

0.9139 (0.8692,0.9587)

0.9042 (0.8913,0.9171)

0.8848 (0.8378,0.9318)

 

SDAFS

0.8556 (0.8389,0.8722)

0.7410 (0.6738,0.8082)

0.9308 (0.9088,0.9527)

0.8958 (0.8776,0.9139)

 

MIFS

0.8533 (0.8368,0.8698)

0.8312 (0.7372,0.9251)

0.9314 (0.9159,0.9467)

0.9027 (0.8740,0.9314)

 

GA

0.8741 (0.8556,0.8926)

0.9557 (0.9121,0.9993)

0.9198 (0.9017,0.9379)

0.9223 (0.8890,0.9568)

BPNN

PCA

0.8493 (0.7781,0.9204)

0.7500 (0.6490,0.8510)

0.9321 (0.9056,0.9585)

0.8177 (0.7701,0.8652)

 

LDA

0.7046 (0.6461,0.7631)

NaN

NaN

NaN

 

MMC

0.8277 (0.7852,0.8702)

NaN

0.8407 (0.7881,0.8932)

0.7505 (0.6789,0.8221)

 

SDAFS

NaN

NaN

NaN

NaN

 

MIFS

0.8683 (0.8302,0.9064)

NaN

NaN

0.7935 (0.6939,0.8932)

 

GA

NaN

NaN

NaN

NaN

  1. The reduced dimensionality for feature extraction algorithm is 15, while the dimensionality for feature selection is 20. The best performance combination for each class and each classifier is displayed in bold.