Columns:
1) Identifier:
--> 2XXX Dateset ID (2=MMSys, 3=UMA, 4=MiBShar, 5=SisFall)
--> X001 Subject number
2) Dataset: (redundant to 1)
(2=MMSys, 3=UMA, 4=MiBShar, 5=SisFall)
3) Subject ID: (redundant to 1)
4) SensorPosition:
Sensor position of the data in the test set.
CHEST, THIGH, WAIST, unknown (=mixed)
5) TrainingDataPosition:
Sensor position of the data in the training set
6) TrainingDataEqualsSensorPosition:
0: Training and test sensor position is not the same
1: Training and test sensor position is the same
7) Group:
Group number that this subject is assigned to based on the clustering
8) Test:
Baseline - Training set: All subjects, all positions; Test set: All positions from one subject
BaselinePosition - Training set: All subjects, all positions; Test set: One position for one subject
Group - Training set: All subjects from same group, all positions; Test set: All positions from one subject
GroupPosition - Training set: All subjects from same group, all positions; Test set: One position for one subject
PositionAndGroupAware - Training set: All subjects from same group, one position; Test set: One position for one subject
PositionError - Training set: One position from all subjects; Test set: One position for one subject
9) Classifier:
The names stand for the following algorithms used:
ANNRangeAlgorithm -> ANN with feature vector 1
ANNRangeAlgorithmGravity -> ANN with feature vector 2
J48FallDetectionAlgorithm -> J48 decision tree with feature vector 3
J48FallDetectionAlgorithmGravity -> J48 decision tree with feature vector 4
J48SumVectorMagAlgorithm -> J48 decision tree with feature vector 5
KNNAlgorithm --> KNN with feature vector 3
KNNAlgorithmGravity -> KNN with feature vector 4
RandomForestAlgorithm -> Random Forest with feature vector 3
RandomForestAlgorithmGravity -> Random Forest with feature vector 4
SisFallAccuracyThresholdAlgorithm -> Threshold algorithm with feature vector 6, optimized for accuracy fall
SisFallSensitivityThresholdAlgorithm -> Threshold algorithm with feature vector 6, optimized for sensitivity fall
SVMKauAndChen2014Algorithm -> SVM with feature vector 7
Feature Vector 1:
- Acceleration range
Feature vector 2:
- Acceleration range
- Gravity
Feature vector 3:
- Mean accleration
- Variance in acceleration
- Standard deviation in acceleration
- Median of acceleration
- Interquartile range
- Mean absolute deviation
- Kurtosis
- Correlation coefﬁcient
- Entropy (Time)
- Energy
- Standard deviation magnitude horizontal
- Sum vector magnitude
- Acceleration range
Feature vector 4:
- Mean accleration
- Variance in acceleration
- Standard deviation in acceleration
- Median of acceleration
- Interquartile range
- Mean absolute deviation
- Kurtosis
- Correlation coefﬁcient
- Entropy (Time)
- Energy
- Standard deviation magnitude horizontal
- Sum vector magnitude
- Acceleration range
- Gravity
Feature vector 5:
- Sum vector magnitude
Feature vector 6:
- Standard deviation magnitude horizontal
Feature vector 7:
- Standard deviation in acceleration
- Sum vectore magnitude
10) CM_0_0:
Confusion matrix at row 0, column 0
11) CM_0_1:
Confusion matrix at row 0, column 1
12) CM_1_0:
Confusion matrix at row 1, column 1
13) CM_1_1:
Confusion matrix at row 1, column 1
14) PrecisionFall
15) RecallFall
16) fMeasureFall
17) PrecisionADL
18) RecallADL
19) fMeasureADL