W14_TRI_ Gold Price Forecast Model for 2012-2016

1.      Problem recognition, definition and evaluation

In previous chapter, the model was selected using regression with Polynomial Formula. Before applying the resulted formula into forecast model, it has to be tested analysis to describe the reliability of the predictive model, thus it can be used for forecast purpose.

2.      Development of the feasible alternative

The model will be tested using residual (error) analysis to assess the quality of a regression.  After it has successfully passed, the formula of gold price model will be used to forecast future gold price.

3.      Development of the outcome

Table 1 and Figure 1 present residual analysis and plot using control chart after comparing Equation output to actual data.

Table 1: Chart Variables for Residual Analysis (source: author)

 

Figure 1: Residual Analysis Chart for Gold Price Forecast Model (source: author)

4.      Criteria of Selection

The model is reliable if it follows certain rules particularly for consistency condition.  

5.      Analysis

This indicates that the model is quite reliable, supported by the following conditions:

a. The variance of residual data is relatively constant - either increasing or decreasing over the period (stationary condition)

b. Residuals show no drift with time (consistency condition)

c. Distribution of residual (error) values show 97.4% are within 2 sigma limit and 98.7% within 3 sigma limit as shown in Figure 2.

Therefore, it will be used as predictive model of future gold price.

Figure 2: Distribution Analysis of Residual Values (source: author)

6.      Selection

Lastly, the range estimate of future gold price with 90% confidence until Dec-2016 is presented in Figure 3 using such tested model. Estimate was done based on historical data and error analysis using its standard deviation (1, 2 and 3 sigma). As result of that, the gold price projected for Dec-2016 is within range of USD 2794 and USD 3207.

Figure 3: Average Monthly Gold Price Forecast Output 2012 –2016 (source: author)

7.      Performance monitoring and post-evaluation of results

The model was well tested, and therefore used to forecast gold price and its associated variables such as steel prices and lending rate for 2012 - 2016. However, the model still needs to be updated since the process outside which builds the model (paper currency system) is destroyed. But, it does not break the fact that gold is the most reliable and sustainable currency.

References:

·         World Gold Council. (2012, Sep 24). Interactive gold price chart and downloads. Retrieved from: http://www.gold.org/download/value/stats/statistics/xls/gold_prices.xls

·         Origin Lab. Graphic Residual Analysis. (2012, Oct 8). Retrieved from: http://www.originlab.com/www/helponline/origin/en/UserGuide/Graphic_Residual_Analysis.html

·         Arsham, H. (2011). Predictions by Regression. Retrieve from: http://home.ubalt.edu/ntsbarsh/stat-data/forecast.htm#rcomputeodel