Forecasting on tuberculosis’s patient at Hospital Raja Perempuan Zainab II from january 2004 until december 2012 using time series analysis / Siti Dawama Izzati Mohammed

In recent years, Tuberculosis or TB is known as one of the deadly disease in the world. Report from Ministry of Health shows that the number of TB was gradually increase for the past twenty. TB is caused by bacteria Mycobacterium and can spread through lymph node and bloodstream to any organ in the body. TB can be either active or latent and this bacterial infection that can most affect lungs. This study is conducted to study the pattern of the total number of Tuberculosis’s patient, to determine the best fitted forecasting model for the Tuberculosis’ patients and to forecast the total number of Tuberculosis’ patients for January 2013 till December 2013 in HRPZ II at Kota Bharu, Kelantan. The data is collected from January 2004 to December 2012. The data were analyzed using Univariate Modelling and Box-Jenkin Methodology. The result shows that the pattern of the total number of TB patients based on time series plot and trend line have a downward trend over the period between 2004 until 2012. The result shows the best model for Univariate Modelling Technique are Single Exponential Smoothing Technique and Holt’s Method since it have smallest MSE and MAPE compared to other model which are 61.78 and 29.57%. For Box-Jenkin Methodology, the best model is ARIMA (1, 1, 1) with MSE 65.90 and MAPE 29.48%. Based on both forecasting modeling technique, it was found that the Univariate Modelling Technique is the best model which Single Exponential Smoothing and Holt’s Method were chosen with the smallest MSE and MAPE which is 61.78 and 29.57%. The forecast value is estimate using Single Exponential Smoothing method since it easy to compute and much simple than Holt’s Method for January 2013 to December 2013 was 18 patients