CUP classification based on a tree structure with MiRNA feature selection

Xiaoxue Zhang, Dunwei Wen, Ke Wang, Yinan Yang

    Research output: Chapter in Book/Report/Conference proceedingPublished Conference contributionpeer-review

    Abstract

    Given the low sensitivity of identifying the origin of cancer tissues using miRNAs in previous research, we adopt a decision tree structure to build a new SVM based model for identifying a variety of Cancer of Unknown Primary Origin (CUP). We use an information gain based feature selection method provided by Weka to select miRNAs and combine them with previously recognized features to determine several most useful miRNAs. Next we design a layer-by-layer classification tree based on the expression levels of these selected miRNAs. Then we use a polynomial kernel SVM classifier, which is more effective in dealing with binary classification problem, for classification at each node of the decision tree structure. In our experiments, a final overall sensitivity of the test set reached 87%, and the sensitivity of identifying the metastatic samples in the test set significantly increased by 9%. The 10-fold cross-validation on this model shows that the sensitivity of the test set is not less than the sensitivity of the training set, indicating that the model has good generalization ability. Additionally, the use of general feature selection makes the approach of this paper more adaptable and suitable for other areas.

    Original languageEnglish
    Title of host publicationAdvances in Artificial Intelligence and Its Applications - 12th Mexican International Conference on Artificial Intelligence, MICAI 2013, Proceedings
    Pages485-496
    Number of pages12
    EditionPART 1
    DOIs
    Publication statusPublished - 2013
    Event12th Mexican International Conference on Artificial Intelligence, MICAI 2013 - Mexico City, Mexico
    Duration: 24 Nov. 201330 Nov. 2013

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    NumberPART 1
    Volume8265 LNAI
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference12th Mexican International Conference on Artificial Intelligence, MICAI 2013
    Country/TerritoryMexico
    CityMexico City
    Period24/11/1330/11/13

    Keywords

    • CUP
    • Feature selection
    • miRNAs
    • Sensitivity
    • SVM

    Fingerprint

    Dive into the research topics of 'CUP classification based on a tree structure with MiRNA feature selection'. Together they form a unique fingerprint.

    Cite this