Forecast Model for Price of Gold: Multiple Linear Regression with Principal Component Analysis


  • Jyothi Manoj Department of Statistics, Kristu Jayanti College, Bangalore, Karnataka, India
  • Suresh K K Department of Statistics, Bharathiar University, Coimbatore, Tamil Nadu, India


Gold price, forecasting, multiple linear regression, multicollinearity, principal component analysis, orthogonal components, residual analysis


The forecast price of gold is significant due to its ever-increasing demand. The price of gold in India is influenced by many other financial variables. These financial variables can be used to predict the price of gold using multiple linear regression method (MLR). The presence of multicollinearity of the explanatory variables is against the assumptions of classical linear regression models; hence to get rid of this problem, principal component analysis (PCA) is carried out to bring in linear combinations of the variables which are correlated. An attempt to develop a regression model to predict the price of gold using seven variables-viz., demand, USD to INR, S&P, NIFTY, BSESENSEX, oil price and US dollar index is made which proved high multicollinearity. Hence, two orthogonal factors are derived using the explanatory variables. This approach improved the prediction accuracy of the model. The coefficient of determination improved to 0.625 from 0.572. Moreover, the analysis which involved seven variables, reduced to two uncorrelated factors will make the interpretation easier. The model is inferred as significant using ANOVA test. Residual analysis also favours the model.


Download data is not yet available.


Gujarati D, Porter DC, Gunasekhar S. Basic econometrics. 5th edition, Special Indian Edition, MacGrawHill Publications; 2012.

Ismail Z, Yahay A, Shabri A. Forecasting gold prices using multiple linear regression method. Am J Appl. Sci. 2009; DOI: 10.3844/ajassp.1509.1514.

Saravanan S, Kannan S, Thangaraj C. India’s electricity demand forecast using regression analysis and artificial neural networks based on principal components. ICTACT J Soft Comput. 2012; 2: 365-370.

Sopipan N, Kanjanavajee W, Sattayatham P. Forecasting SET50 index with multiple regression based on principal component analysis. J App Finance Bank. 2012; 2: 271-294.

Toraman C, Basarir C, Bayramoglu MF. Determination of factors affecting the price of gold: a study of MGARCH model. Bus Econ Res J. 2011; 2: 1309-2448.

Ul-Saufie AZ, Yahya AS, Ramli NA. Improving multiple linear regression model using principal component analysis for predicting PM10 concentration in Seberang Prai, Pulau Pinang. Int J Env Sci. 2011; 2: 403-410.