Extracting Partial Association Mantel-Haenszel Test based Cause and Effect Relationships using Decision Tree

  • Unique Paper ID: 161242
  • Volume: 10
  • Issue: 3
  • PageNo: 58-63
  • Abstract:
  • Decision trees are very useful tools for both data classification and regression in many real time situations in data mining, machine learning, big data analytics including many distributed data applications. Usage of a single data analysis technique is common in many machine learning techniques. Many recent trends are being becoming popular towards the usage of hybrid techniques in data analytics. As a result, the standard benchmarking decision tree classifiers are combined with many other statistical techniques in order to find cause and effect relationships present in the given datasets. Causal relationships are computed between a predictor (input) variable and the outcome (output) variable. One of the most popular statistical based data analysis techniques called Partial Association Mantel-Haenszel Test combined with bench mark decision tree classifiers in order to elucidate cause and effect relationships in the datasets. These hybrid techniques are applied on the simple and hypothetical dataset for finding cause and effect relationships in the dataset. Experiments are conducted on the selected dataset, customers. Experimental results have revealed that the identified cause and effect relationships present in the dataset are real and well-matched with many of the real time situational scenarios.

Related Articles