R. Setiono, Generating concise and accurate classification rules for breast cancer diagnosis, Artificial Intelligence in Medicine 18 (2000) 205--219.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....can predict breast cancer from the Wisconsin dataset. He needed first to train an ANN using BP and achieved an accuracy level on the test data of approximately 94 . After applying his rule extraction technique, the accuracy of the extracted rule set did not change. In a more recent work, Setiono [27] used feature selection before training the ANN. The new rule sets had an average accuracy of more than 96 . This is an improvement when compared to the initial results. It is also comparable to the results of Fogel et al. 8] Furundzic et al. 10] presented another BP ANN attempt where they ....

R. Setiono. Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18:205--219, 2000. 21

....2 classes and 9 attributes (Class 1: 1 445, Class 2: 446 683) Table 1: Classification rates and model complexity for classifiers constructed for the Wisconsin Breast Cancer problem. # denotes results from averaging a ten fold validation. Author Method # Rules # Conditions Accuracy Setiono [21] NeuroRule 1f 4 4 97.36 Setiono [21] NeuroRule 2a 3 11 98.1 Pena Reyes and Sipper [16] Fuzzy GA1 1 4 97.07 Pena Reyes and Sipper [16] Fuzzy GA2 3 16 97.36 Nauck and Kruse [15] NEFCLASS 2 10 12 95.06 # GA and fuzzy logic were also applied to this problem [16] In this method the number of ....

....1: 1 445, Class 2: 446 683) Table 1: Classification rates and model complexity for classifiers constructed for the Wisconsin Breast Cancer problem. # denotes results from averaging a ten fold validation. Author Method # Rules # Conditions Accuracy Setiono [21] NeuroRule 1f 4 4 97.36 Setiono [21] NeuroRule 2a 3 11 98.1 Pena Reyes and Sipper [16] Fuzzy GA1 1 4 97.07 Pena Reyes and Sipper [16] Fuzzy GA2 3 16 97.36 Nauck and Kruse [15] NEFCLASS 2 10 12 95.06 # GA and fuzzy logic were also applied to this problem [16] In this method the number of rules to be generated needs to be ....

[Article contains additional citation context not shown here]

Setiono R. (2000) Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18: 205--219.

....either on accuracy or interpretability. Recently some approaches to combining these properties have been reported; fuzzy clustering is proposed to derive transparent models in [10] linguistic constraints are applied to fuzzy modeling in [9] and rule extraction from neural networks is described in [11]. Hence, to obtain compact and interpretable fuzzy models model reduction algorithms have to be used that will be overviewed in Section 4. Search method: consists of two components: parameter search and model search. Once the model representation and the model evaluation criteria are fixed, then ....

....either on accuracy or interpretability. Recently some approaches to combining these properties have been reported; fuzzy clustering is proposed to derive transparent models in [10] linguistic constraints are applied to fuzzy modeling in [9] and rule extraction from neural networks is described in [11]. 4.1 Similarity driven rule base simplification The similarity driven rule base simplification method [22] uses a similarity measure to quantify the redundancy among the fuzzy sets in the rule base. A similarity measure based on the set theoretic operations of intersection and union is ....

R. Setiono, Generating concise and accurate classification rules for breast cancer diagnosis, Artificial Intelligence in Medicine 18 (2000) 2054 19.

....detail in Section 2.2, I concentrate on the former. Three main streams can be identified in the research on hybrid neural fuzzy systems: Fuzzy rule extraction from neural networks. This approach attempts to extract, in the form of fuzzy rules, the knowledge embedded in trained neural networks [32, 107, 161]. The main drawback of these techniques is that the access to the knowledge requires a previous rule extraction phase. Neuro fuzzy systems. These are fuzzy inference systems implemented as neural networks, taking advantage of their structural similarity (see Section 1.2.4) The main advantage ....

....found systems whose performance exceeds 98.5 . these results are summarized in Figure 4.1. Table 4.3 compares the best systems found by Fuzzy CoCo with the top systems obtained by the fuzzy genetic approach (Section 2. 4) 125, 130] and with the systems obtained by Setiono s NeuroRule approach [161] (note that the results presented by these two works were the best reported to date for genetic fuzzy and neuro Boolean rule systems, respectively, and that they were compared with other previous approaches such as [160,162,169] The evolved 94 95 96 97 98 99 0 40 80 120 160 Classification ....

[Article contains additional citation context not shown here]

R. Setiono. Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18(3):205 -- 219, 2000.

....and 16 of these are omitted because these are incomplete, which is common with other studies. The class distribution is 65.5 benign and 34.5 malignant, respectively. This classification problem is widely used to test the effectiveness of classification and rule extraction algorithms. Setiono [4] applied a feedforward neural network on preprocessed data followed by a two step rule extraction. Nauck and Kruse [2] combined neuro fuzzy techniques with interactive strategies for rule pruning to obtain a fuzzy classifier. An initial rule base was made by applying two sets for each input, ....

....strategies for rule pruning to obtain a fuzzy classifier. An initial rule base was made by applying two sets for each input, resulting in 2 =512 Table 1. Classification rates on the Wisconsin Breast Cancer data. Method Best result Aver result Worst result Rules Model eval Corcoran and Sen [4] Ischibuchi et al. Nauck and Kruse [2] This paper feature 2 = 2 : feature 6 = 3 : 1 feature 6 3 : feature 1 = 3 : 1 feature 1 3 : 2 feature 2 2 : feature 2 4 : 2 feature 2 = 4 : feature 6 2 : 2 feature 6 = 2 : ....

Setiono R. (2000) Generating concise and accurate classification rules for breast cancer diagnosis 18, 205-219.

.... medical tasks [16] they have not yet been widely accepted in medicine [15] Fommately, during the last decade much work has addressed the issue of improving the comprehensibility of artificial neural networks [ 1 ] 30] and some results have already been applied to medical tasks [13] 25] [26]. Artificial neural network ensemble is a learning technique where multiple artificial neural networks are trained to solve the same problem. Since the generalization ability of learning systems based on artificial neural networks can be significantly improved with this technique, it has become a ....

R. Setiono, "Generating concise and accurate classification rules for breast cancer diagnosis," Artificial Intelligence in Medicine, vol. 18, no.3, pp.205-219, 2000.

....methods that combine GA and fuzzy logic were also applied to this problem [24] In this method the number of rules to be generated needs to be determined a priori. This method constructs a fuzzy model that has four membership functions and one rule with an additional else part. Setiono [25] has generated similar compact classifier by a two step rule extraction from a feedforward neural network trained on preprocessed data. As Tab. 1 shows, our fuzzy rule based classifier is one of the most compact models in the literature with such high accuracy. 5.3 Example 3: Comparative Study ....

....classification problem to present how the performance and 24 Classification rates and model complexity for classifiers constructed for the Wisconsin Breast Cancer problem. # denotes results from averaging a ten fold validation. Author Method # Rules # Conditions Accuracy Setiono [25] NeuroRule 1e 1 4 97.36 Setiono [25] NeuroRule 1f 4 4 97.36 Setiono [25] NeuroRule 2a 3 11 98.1 Pena Reyes and Sipper [24] Fuzzy GA1 1 4 97.07 Pena Reyes and Sipper [24] Fuzzy GA2 3 16 97.36 Nauck and Kruse [5] NEFCLASS 2 10 12 95.06 # This paper DT based FC 2 3 4 96.82 # the ....

[Article contains additional citation context not shown here]

R. Setiono, Generating concise and accurate classification rules for breast cancer diagnosis, Artificial Intelligence in Medicine 18 (2000) 205--219.

....little understandability, i.e. diagnostic decisions are essentially black boxes, with no explanation as to how they were attained. With increased interpretability in mind as a prior objective, a number of researchers have applied the method of extracting Boolean rules from neural networks [38, 39]. Their results are encouraging, exhibiting both good performance and a reduced number of rules and relevant input variables. Nevertheless, these systems use Boolean rules and are not capable of furnishing the user with a measure of confidence for the decision made. Our own work on the evolution ....

....led to a fuzzy system whose performance exceeds 98.0 , and of these, 81 runs found systems whose performance exceeds 98.5 . Table 6 compares our best systems with the top systems obtained by the fuzzy genetic approach (Section 4) 26] and with the systems obtained by Setiono s NeuroRule approach [38]. The evolved fuzzy systems described in this paper can be seen to surpass those obtained by other approaches in terms of performance, while still containing simple, interpretable rules. As shown in Table 6, we obtained higher performance systems for all rule base sizes but one, i.e. from ....

[Article contains additional citation context not shown here]

R. Setiono. Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18(3):205 -- 219, 2000.

....5 1 1 . 1 benign 2 5 4 4 . 1 benign : 683 4 8 8 . 1 malignant There are several studies based on this database. Among them, researchers having interpretability of the diagnostic as a prior objective, have applied the method of extracting Boolean rules from neural networks [44, 47]. Our own work on the evolution of fuzzy rules for the WBCD problem has shown that it is possible to obtain diagnostic systems exhibiting high performance, coupled with interpretability and a confidence measure [27 30] In our previous work we used a simple genetic algorithm rather than Fuzzy ....

....rate) 241 runs led to a fuzzy system whose performance exceeds 98.0 , and of these, 81 runs found systems whose performance exceeds 98.5 . Table 8 compares our best systems with the top systems obtained in our previous work [29] and with the systems obtained by Setiono s NeuroRule approach [44] (note that Pena Reyes Sipper, Fuzzy Modeling by Fuzzy CoCo 25 Database SL SW PL PW P 1 4.65 2.68 4.68 0.39 P 2 4.65 3.74 5.26 1.16 P 3 5.81 4.61 6.03 2.03 Rule base Rule 1 if (PW is Low) then (setosa is Yes) versicolor is No) virginica is No) Rule 2 if (PL is Low) and (PW is ....

[Article contains additional citation context not shown here]

R. Setiono. Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18(3):205 -- 219, 2000.

....rate) 241 runs led to a fuzzy system whose performance exceeds 98.0 , and of these, 81 runs found systems whose performance exceeds 98.5 . Table IV compares our best systems with the top systems obtained in our previous work [26] and with the systems obtained by Setiono s NeuroRule approach [37] (note that the results presented by these two works were the best reported to date for genetic fuzzy and neuro Boolean rule systems, respectively, and that they were compared with other previous approaches such as [36,38,39] The evolved fuzzy systems described in this paper can be seen to ....

....seven rule systems, while all our one rule systems perform as good as the best system reported by Setiono. TABLE IV Comparison of the best systems evolved by Fuzzy CoCo with the top systems obtained using single population evolution [26] and with those obtained by Setiono s NeuroRule approach [37]. Shown below are the classification performance values of the top systems obtained by these approaches, along with the number of variables of the longest rule in parentheses. Results are divided into seven classes, in accordance with the number of rules per system, going from one rule ....

[Article contains additional citation context not shown here]

R. Setiono. Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine, 18(3), 2000.

....either on accuracy or interpretability. Recently some approaches to combining these properties have been reported; fuzzy clustering is proposed to derive transparent models in [9] linguistic constraints are applied to fuzzy modeling in [13] and rule extraction from neural networks is described in [8]. In this paper we describe an approach that addresses both issues. Compact, accurate and linguisticly interpretable fuzzy rule based classifiers are obtained from labeled observation data in an iterative fashion. An initial model is derived from the observation data and subsequently, feature ....

Setiono R. (2000) Generating concise and accurate classification rules for breast cancer diagnosis. Artificial Intelligence in Medicine 18, 205-219.

No context found.

R. Setiono, "Generating concise and accurate classification rules for breast cancer diagnosis," Artif. Intell. Med., vol. 18, pp. 205--219, 2000.

No context found.

R. Setiono, "Generating concise and accurate classification rules for breast cancer diagnosis," Artif. Intell. Med., vol. 18, pp. 205--219, 2000.

No context found.

R. Setiono, Generating concise and accurate classification rules for breast cancer diagnosis, Artificial Intelligence in Medicine 18 (2000) 205--219.

No context found.

R. Setiono, Generating concise and accurate classification rules for breast cancer diagnosis, Artificial Intelligence in Medicine 18 (2000) 205--219.

No context found.

R. Setiono, "Generating concise and accurate classification rules for breast cancer diagnosis, " Artificial Intelligence in Medicine, vol. 18, pp. 205--219, 2000.

No context found.

Setiono,R.; "Generating concise and accurate classification rules for breast cancer diagnosis "; Artificial Intelligence in medicine 18:205-219, 2000

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC