Takagi-Sugeno Fuzzy Interference Systems Analysis
4. ANFIS: On the basis of Takagi-Sugeno Fuzzy interference system the ANFIS or the Adaptive Network Based Fuzzy Inference is one kind of AI or Artificial Intelligence. In the 1990s this technique was developed as it combines the two forms the Fuzzy Logic Principles and as well as the Neural Networks. It can gather both the framework. The interference system correlates to the set of the IF-THEN functions which are non linear to be approximated using this rules it has a capability to learn. Thus it is considered to be the Universal Estimator. When genetic algorithm gathers the best ranged information it is more efficient while using the ANFIS. [26]
Now that all the variables to be used were decided the next step required the model to be created. …show more content…
As can be seen with the first two columns list the ICD-O and the frequency, i.e. the number of times that ICD-O has been diagnosed at NICRH. The next three columns list the demographic attributes and should be considered as independent events. The output column titled as Risk of being diagnosed is a probability calculation of the combination of events for each set of inputs. In order to provide a better explanation of the process, consider the first input from Table 2 as an example.
Here we need to measure the Risk of being diagnosed with C34 where the patient is an illiterateagri-worker who smokes tobacco.
Therefore,
p(Risk of an illiterate agri-worker who smokes tobacco being diagnosed with C34) = p(Patient is diagnosed with C34) * p(Patient is a Smoker) * p(patient is an agri-worker) * p(Patient is illiterate)
Thus the measured probability of these set of independent events is 0.02168, which is set at the output in the dataset …show more content…
The FIS for males is shown in figure 2 and it is similar to the FIS for females. The subcategories for each attribute as shown in Figure 1, is taken as members for the corresponding fuzzy variable. The detailed depiction of these membership functions are shown in Figure 5. The members of most input variables remain the same for the FIS of both males and females, with the exception of the input variables ICD-O and occupation. There exist two membership structures for the input occupation due to the inclusion of the attribute housewife for females as shown in 5e and 5f.As for the membership function ICD-O for male and female shown in 5a and 5b, the structure differs as the top five malignancies for each gender is different in
Now that all the variables to be used were decided the next step required the model to be created. …show more content…
As can be seen with the first two columns list the ICD-O and the frequency, i.e. the number of times that ICD-O has been diagnosed at NICRH. The next three columns list the demographic attributes and should be considered as independent events. The output column titled as Risk of being diagnosed is a probability calculation of the combination of events for each set of inputs. In order to provide a better explanation of the process, consider the first input from Table 2 as an example.
Here we need to measure the Risk of being diagnosed with C34 where the patient is an illiterateagri-worker who smokes tobacco.
Therefore,
p(Risk of an illiterate agri-worker who smokes tobacco being diagnosed with C34) = p(Patient is diagnosed with C34) * p(Patient is a Smoker) * p(patient is an agri-worker) * p(Patient is illiterate)
Thus the measured probability of these set of independent events is 0.02168, which is set at the output in the dataset …show more content…
The FIS for males is shown in figure 2 and it is similar to the FIS for females. The subcategories for each attribute as shown in Figure 1, is taken as members for the corresponding fuzzy variable. The detailed depiction of these membership functions are shown in Figure 5. The members of most input variables remain the same for the FIS of both males and females, with the exception of the input variables ICD-O and occupation. There exist two membership structures for the input occupation due to the inclusion of the attribute housewife for females as shown in 5e and 5f.As for the membership function ICD-O for male and female shown in 5a and 5b, the structure differs as the top five malignancies for each gender is different in