Introduction
Methods
Patient Characteristics
Image Acquisition
Tumor Segmentation
Feature Selection
Feature selection method | Definition |
---|---|
Cfs Subset Eval (CF SUB E) | Evaluates the worth of a subset of attributes by considering the individual predictive ability of each feature along with the degree of redundancy between them. |
Correlation Attribute Eval (CO AT EV) | Evaluates the worth of an attribute by measuring the correlation (Pearson’s) between it and the class. |
Gain Ratio Attribute Eval (GA FA AT) | Evaluates the worth of an attribute by measuring the gain ratio with respect to the class. |
One R Attribute Eval (One R AT) | Evaluates the worth of an attribute by using the One R classifier. |
Relief F Attribute Eval (RE F AT) | Evaluates the worth of an attribute by repeatedly sampling an instance and considering the value of the given attribute for the nearest instance of the same and different class. |
Symmetrical Uncert Attribute Eval (SYM AT) | Evaluates the worth of an attribute by measuring the symmetrical uncertainty with respect to the class. |
Result
Patients and Response
Demographics | Frequency N | Percent % |
---|---|---|
Gender | ||
Male | 44 | 65.7 |
Female | 23 | 34.3 |
Total | 67 | 100 |
Age | ||
18–40 | 15 | 22.4 |
41–60 | 23 | 34.3 |
> 61 | 29 | 43.3 |
Total | 67 | 100 |
Response | ||
Grade 0 | 11 | 9.4 |
Grade 1 | 19 | 26.4 |
Grade 2 | 26 | 47.2 |
Grade 3 | 11 | 17 |
Texture Analysis
Classifiers | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
Bayesian network | Naive bayesian network | Ada boost M1 | Iterative classifier optimizer | Logit boost | Randomizable filtered classifier | Random sub space | Random forest | K logistic model tree | ||
Filter value | ||||||||||
T1W | No filtration | 63.9 | 51.2 | 50.0 | 51.0 | 51.2 | 50.0 | 52.2 | 51.3 | 50.0 |
0.5 (fine) | 52.3 | 54.2 | 56.8 | 58.1 | 53.8 | 53.2 | 59.1 | 55.3 | 50.0 | |
1 (medium) | 56.2 | 68.0 | 78.1 | 70.3 | 78.1 | 52.5 | 79.3 | 53.1 | 52.9 | |
1.5 (coarse) | 52.1 | 50.0 | 51.2 | 51.2 | 52.1 | 68 | 51.3 | 50.0 | 50.0 | |
T2W | Filter value | |||||||||
No filtration | 66.7 | 51.1 | 74.8 | 71.8 | 63.4 | 51.5 | 60.5 | 54.3 | 51.1 | |
0.5 (fine) | 72.5 | 85.1 | 79.4 | 81.3 | 79.3 | 66.2 | 72.3 | 71.8 | 61.5 | |
1 (medium) | 57.6 | 65.0 | 58.9 | 58.6 | 60.2 | 68.8 | 55.2 | 56.2 | 50.0 | |
1.5 (coarse) | 57.6 | 80.6 | 80.1 | 53.6 | 80.1 | 67.1 | 66.4 | 57 | 60.9 |
Feature Selection Performance
Classifiers | |||||||||
---|---|---|---|---|---|---|---|---|---|
Feature selection | Bayesian network | Naive Bayesian network | Adaboost M1 | Iterative classifier optimizer | Logit Boost | Randomizable filtered classifier | Random sub space | Random forest | K logistic model tree |
Cfs Subset Eval | 72.5 | 85.1 | 79.4 | 81.3 | 79.3 | 66.2 | 72.3 | 71.8 | 77.5 |
Correlation Attribute Eval | 61.2 | 57.8 | 73.1 | 55.9 | 75.6 | 72.0 | 55.7 | 51.6 | 58.8 |
Gain Ratio Attribute Eval | 56.2 | 52.7 | 73.1 | 58.4 | 74.8 | 62.6 | 55.7 | 52.0 | 59.7 |
One R Attribute Eval | 56.2 | 55.1 | 73.9 | 59.0 | 74.8 | 64.1 | 50.2 | 55.1 | 58.8 |
Relief F Attribute Eval | 57.3 | 52.7 | 72.0 | 54.1 | 75.9 | 59.3 | 51.6 | 52.7 | 58.8 |
Symmetrical Uncert Attribute Eval | 65.3 | 54.8 | 73.2 | 55.3 | 74.5 | 66.4 | 56.5 | 55.4 | 66.8 |