¹Department of Information Technology, Bahauddin Zakariya University, Pakistan.
²Department of Computer Science and Information Technology, Women University, Pakistan.

---

Abstract—Various kinds of pressure and unbalanced eating behaviors, along with alcohol inhalation and on-going toxic gases, absorption of tainted nutrients, unnecessary intake of cured food and ingestion of drug, enables patients to increase year by year from liver disease. For this purpose, the type of data mining algorithms can help medical doctors to diagnose patients in hospital. This paper analyzes meta learning algorithms to classify the Indian liver patient dataset. The Data set is attained from UCI repository that contains 583 instances. Adaboost, logitboost, Bagging and Grading meta learning algorithms are applied to this data set. These algorithms are compared on the basis of Correct Classification, Incorrect Classification and Time to build model. Grading is the best algorithm among these meta learning algorithms as it provides highest Correct Classification rate and minimum rate of incorrect classification. Execution time for Grading is less than Adaboost, Logitboost and Bagging. Key role is played by Grading algorithm in shaping enhanced classification accuracy (Correct Classification Rate) of a data set.

Index Terms—Disease diagnostic, data mining, machine learning, meta learning.

[PDF]

Cite: Maruf Pasha, Meherwar Fatima, "Comparative Analysis of Meta Learning Algorithms for Liver Disease Detection," Journal of Software vol. 12, no. 12, pp. 923-933, 2017.