Multi-features Based Arrhythmia Diagnosis Algorithm Using Xgboost

Arrhythmia is the common disease in today's society. In order to judge the specific situation of the patient, doctors often observe the ECG (Electrocardiograph) signal to get enough information to help them diagnose. Many researches have been devoted to using various machine learning algorithms to study the classification problem of arrhythmia. However, single judgement cannot achieve acceptable results. Obviously, electrocardiogram signal data set has many missing data for various reasons. Fortunately, Xgboost has a strong ability to handle these problems. In this paper, the Multi-features Arrhythmia Diagnosis Algorithm Xgboost (MADA-Xgboost) is used in classification of arrhythmia. The data in the dataset is firstly divided into discrete data and continuous data. Cluster analysis algorithm is used to discretize continuous data. After that, integrate the processed data with the original discrete data. Then use principal components analysis (PCA) to make feature judgments and select the appropriate features. The next step is generating features vector through the chosen features. If the vector is generated, it can be used in Xgboost to adjust parameters and train the model. The modified parameters are number of iterations, depth, learning rate, alpha and gamma. Finally, after getting the trained model, using it to process the test data and classify the data as normal or abnormal. In order to prove the effectiveness of MADA-Xgboost, we analyze the data and algorithm. Obviously, people of different ages, genders, heights and weights have different health status. These data will affect the diagnosis of heart disease. At the same time, systolic blood pressure and diastolic blood pressure are also associated with the human health. In addition, Xgboost is not only highly accurate but also fast in classification. In this case, the MADA-Xgboost used in the paper is reasonable and correct.