Objective: To provide a basis for the prediction and early diagnosis of laryngeal cancer through data mining and analysis.
Method: We constructed a laryngeal cancer data warehouse on Caché and combined IBM Cognos for on-line analytical processing. After building research data set, software Weka was used for feature selection and data mining.
Result: Data warehouse of laryngeal cancer was structured and Laryngeal cancer data were multidimensional presented. It is possible to find the high relevant and potential characteristics associated with laryngeal cancer by integration and 2-phase feature subset of data mining to extract the larynx.
Conclusion: Application of data mining technology could help clinicians to find potential clinical knowledge and make early diagnosis of laryngeal cancer.