Forecasting free essay sample

The purpose of the project is to determine the most suitable technique to generate the forecast of cocoa production. The models understudied are based on Univariate Modelling Techniques i. e. Naive with Trend Model, Average Change Model, Average Percent Change Model, Single Exponential Smoothing, Double Exponential Smoothing and ARESS method. These models are normally used to determine the short-term forecasts (one month ahead) by analyzing the pattern such as monthly cocoa production. The performances of the models are validated by retaining a portion of the monthly observations as holdout samples. The selection of the most suitable model was indicated by the smallest value of mean square error (MSE) and mean absolute percentage error (MAPE). Based on the analysis, ARRES Method Model is the most suitable model for forecasting monthly cocoa production. Keywords: Univariate Modelling Techniques; Forecast Model; Mean Square Error, Mean Absolute Percentage We refer very frequently to future events in our daily lives, we look forward, we have the foresight to do something, we are able to foretell, we foresee an event and we say that something is forthcoming. We will write a custom essay sample on Forecasting or any similar topic specifically for you Do Not WasteYour Time HIRE WRITER Only 13.90 / page Forecasting can be defined as the science and the art to predict a future event with some degree of accuracy. There are two types of forecast which are event forecast and time series forecast. The future occurrence of an outcome and the timing of such an occurrence are referring to an event forecast. The use of time series information in the prediction of the variable interest is the term of time series forecast. In a time series data set, the information is arranged according to time. Univariate Modelling Techniques are methods for analyzing data on a single variable at a time. Examples of Univariate Modelling Techniques are the Naive Models, Methods of Average, the Exponential Smoothing Techniques and the Box-Jenkins Methodology. Both Double Exponential Smoothing and Holts Method illustrated in this study are classified in the Exponential Smoothing Techniques. Other models available in this same category are Single Exponential Smoothing, Adaptive Response Rate Exponential Smoothing (ARRES), Holts Method and Holt- Winters Trend Seasonality. OBJECTIVE OF THE STUDY The objective of the study is to choose the most suitable model to forecast the cocoa production. The output of the study will serve as a guide in selecting a model for future forecasting or projection of cocoa production. Forecasting on cocoa production can make the supplier easy to understand the demand of cocoa. Cocoa serves as an important crop around the world: a cash crop for growing countries and a key import for processing and consuming countries. Cocoa travels along a global supply chain crossing countries and continents The complex production process involves numerous parties including, farmers, buyers, shipping organizations, processors, chocolates, and distributers. Cultivation of cocoa at the farm level is a delicate process as crops are susceptible to various conditions including weather patterns, diseases, and insects. Unlike larger, industrialized agribusinesses, the vast majority of cocoa still comes from small, family-run farms, who often confront outdated farming practices and limited organizational leverage. A steady demand from worldwide consumers draws numerous global efforts and funds committed to support and improve cocoa farm sustainability. The major producing countries which is in Africa: Cotedlvoire (40% global), Ghana, Nigeria and Cameroon. In Asia and Oceania: Indonesia, Malaysia and Papua New Guinea. In Americas: Brazil, Ecuador and Colombia. In Africa and Asia, a typical farm covers 2 to 5 hectares (4. 9 -12. 3 acres). Small cocoa farms provide more than 90% of world cocoa production. Short-range forecasts of cocoa production are important for the formulation of policy by private concerns in the cocoa trade, by governments and public agencies of cocoa exporting countries, and by the International Cocoa Organization (ICCO). In spite of very large stocks of cocoa held in exporting countries, and notwithstanding the International Cocoa Agreement, prices of cocoa in recent years have responded significantly to major year-to-year changes in production. This response was demonstrated late in 1963 when prices rose sharply with anticipations (later realized) of a record low Brazilian crop for 1964. Again in late 1969 prices rose sharply in response to reports of frost affecting the 1970 Brazilian crop. Executive committee in London illustrates the main changes in production between 2002/2003 and 2011/2012, using a three year moving average to smooth out the effect of weather related aberrations. During this period, world production increased by 3. % per annum. Africas production expanded at an average annual rate of 3. 7% and its share of world production rose. METHODOLOGY In case study, data cocoa productions (tones) were used from year 2003 until 2011 and the data for cocoa production are in monthly. Based on data, the researcher used excels to fit the 6 model. The 6 model is Naive with Trend Model, Simple Exponential Smoothing Model, Double Exponential Smoothing Model, Average Change Model, Average Percent Change Model and ARESS method. For the each model, the initial value had been identified, mean square error (MSE) and mean bsolute percentage error (MAPE). For Simple exponential, double exponential smoothing model and ARRES method, the researcher used try and error method to find the best value of alpha and beta to get the smallest error for the model. From the output, the researcher compared the value of MSE and MAPE for each model to determine which model can be used to forecast the cocoa production. The best model has the smallest value of MSE and MAPE. When developing forecast model, the researcher divide the data set into two parts The tirst part is fitting and t second part is hold-out. Finally the researcher presents all MSE and MAPE in summary table and makes the conclusion the best model that can be used to forecast the cocoa production. There are six models that had been used to find the best model to forecast the production cocoa: i. Naive Model With Trend Naive model is modified to take this characteristic into account. The application of this model is fairly common among organizations. One reason for its popularity is that it can be used even with fairly short time series. Thus overcoming the common problem in most organizations where insufficient data would prohibit the application f sophisticated modeling techniques. Ft+l † Where Ft+l = is the I-step-ahead forecast at period i made in period t for 1=1, 2, 3 = is the actual observation at the time t. it. Average Change Model The average change model is based on the premise that the forecast value is equal to the actual value in the current period plus the average of the absolute changes experienced up to that point in time. This model is useful when the historical data being analyzed are characterized by period-to-period changes that are approximately of the same size. However, this model tends to lag behind turning points and that all periods are weighted equally, irrespective of their importance, when deriving the forecast values.