RESEARCH REVIEW ON COMPARATIVE STUDY OF DIFFERENT FORECASTING METHODS FOR VARIATION IN GOLD PRICE
1Tarun Kumar Yadav, 2Rajul Namdeo
1Asst. Prof., 2 M.Tech Scholar
1, 2 Department of Mechanical Engineering
1, 2 Babulal Tarabai Institute of Research & Technology, Sagar (M.P
.)
, IndiaAbstract : Now a day, the forecasting of various metals i.e. Gold, Silver, Platinum, Nickel, Iron etc. and also forecasting the agricultural product prices i.e. Crop price, Onion price etc. is the critical task. In this review paper, the various researchers used different forecasting techniques i.e. simple moving average, weighted moving average, exponential smoothing method etc are the old methods of forecasting. In today, the researchers used various modern techniques i.e. ARIMA model, Box-Jenkins method, Holter-Winters model, gene expression programming, artificial neutral networks etc. for forecasting. The gold is the precious metal of the world and it is used for making jewelry & other purpose. Thus, the forecasting of gold price is very important for the investor because these forecasted prices help the investors to know suitable time for buying and selling the gold in advance.
Key Words – Forecasting, Gold, ARIMA, Box-Jenkins Methods, Holter-Winters Model etc.
I. INTRODUCTION
Forecasting is defined as, “The process of finding a future demand by using past data or past experience.”
According to Evan J. Douglas, "Demand forecasting may be defined as the process of finding values for demand in future time periods.[21]
Various application of forecasting is as follows:-
Supply Chain Management
Economic forecasting
Earthquake forecasting
Technology forecasting
Weather forecasting, etc.
II. EFFORTBYVARIOUSRESEARCHERS
1. Syazwani Zainal Abidin, Tengku Mardhiah Tengku Jalal, Fadila Amira Razali, Nor Hidayah Hassim and Nur Fatihah Haron (2018), “Comparison on Estimating Malaysia Gold Price via Nonlinear Prediction Method and Box–
Jenkins Model”, studies that the investment of gold increases very greatly in last few year. So, the gold investors to know best suitable model to provide knowledge of current gold price and also, helps the investors to make decision for buy or sell the gold. The aim of this research is to find discontinuous behaviour of gold in form of price changes by using nonlinear prediction (NLP) method and Box-Jenkins method over a period from Jan 4, 2016 to Dec 30, 2016 in Malaysia. Therefore, the forecasting of the gold price was done for the first quarter of year 2017 in Malaysia. In this research Box-Jenkins method consist of three different models based on Autocorrelation function (ACF) and Partial Autocorrelation Function (PACF).
Then these three model were compare with Bayesian Information Criterion (BIC), Mean Square Error (MSE) and Mean Absolute Percentage Error (MAPE) for finding the better model to suitable for time series data of gold price in Malaysia.
The result of this research shows that the ARIMA (1,1,0) was the best model for forecasting the gold price trend in Malaysia.
A prediction performance comparison was done between NLP & Box-Jenkins models on the basis of MAPE and this comparison indicates that NLP provide better prediction performance as compare to Box-Jenkins model.
2. Hesam Dehghani and Mahsa Zangeneh (2018), “Crude Oil Price Forecasting: A Biogeography-Based Optimization Approach”, studies that Crude oil pricing is affected by a variety of factors and is ever changing in international market which has increased the demand for developing accurate models for forecasting prices. This research analyses and tests a time series and biogeography-based optimization (BMMR–BBO) for the estimation of the West Texas Intermediate (WTI) crude oil price. Through the experiments we can safely conclude that BMMR-BBO gives more accurate results then other time series models and since it is more consistent, this model can be used for prediction and analyses of Crude oil prices which in turn will help to form policies related to crude oil and industrial development.
3. Naliniprava Tripathy (2017), “Forecasting Gold Price With Auto Regressive Integrated Moving Average Model”, In this research, ARIMA model was used to forecast the gold price in India over a period of 25 years from July 1990 to Feb 2015. In this research, MAE, RMES, MAPE, MAE and MAPE also use to evaluate the correctness of the model. The result of this research shows that ARIMA (0,1,1) is most suitable model for forecasting the Indian gold price and these gold price helps the investor, market regulator etc to take investment decision.
4. Sneha Chaudhry, Dr. Y. S. Negi, Dr. Rakesh K. Shukla (June 2017), “A Time Series Analysis of Auction Prices of Indian Tea”, Tea is an important commercial crop in India and has been one of the important agricultural export items. In this research, monthly auction price data consist of three year data (Jan 2012- Dec 2014) from tea board of India used to analyze the price behaviour and also two tea growing regions i.e. north and south India for analysis. ARIMA techniques used for this analysis at lag 1. In this research two function i.e. autocorrelation (ACF) and partial autocorrelation function (PCF) used.
5. A Rahman and A S Ahmar (sept.2017), “Forecasting of Primary Energy Consumption Data In The United States: A Comparison Between ARIMA and Holter-Winters Models”, In this research, to compare the ARIMA model and Holt- Winters model based on MAE, RSS, MSE and RMS criteria in forecasting Primary Energy Consumption Total Data in the Us. The data used for this research ranges from Jan 1973 to Dec 2016 and also R-software used for this research. The result of this data analysis is shows that, Holt Winters additive type model is most significant model in forecasting the primary energy consumption total data in the US. This model is more significant when compare with Holt winters multiplicative type and ARIMA seasonal model.
6. Paulo V. Cenas (2017), “Forecast of Agricultural Crop Price using Time Series and Kalman Filter Method”, The aim of this research is to improve the accuracy and correctness of time series model in forecasting future price of rice crop with the combination of ARIMA and Kalman filter techniques respectively. Also five years rice data collected for this research. Then, performance of ARIMA model was compared to the combined performance of ARIMA- Kalman filter by the help of Mean Square Error (MSE) and Root Mean Square Error (RMSE) as the bases of comparison. Result of this research shows that the combined study of ARIMA time series model and Kalman filter technique provide the better and accurate future price of rice.
7. Syed Misbah Uddin, Aminur Rahman, Emtiaz Uddin Ansari (2017), “Comparison of Some Statistical Forecasting Techniques With GMDH Predictor: A Case Study”, the aim of this research to determine the accurate models for forecasting the cement demand. In this research monthly sales data of cement from ranging Jan 2007 to Feb 2016 is used and also Group Method of Data Handling (GMDH) model used to forecasting the cement demand. The time series smoothing techniques such as exponential smoothing, double exponential smoothing, moving average, weightage moving average and regression method were also used in this research. Then, The original data were compared to the forecast produced by the time series model and GMDH model and for comparing the forecasting accuracy mean absolute deviation (MAD), mean absolute percentage error (MAPE) and mean square error (MSE) were also calculated. These comparison shows that the GMDH provide better result than other statistical models based on MAD, MAPE and MSE.
8. Aysenur Budak, Alp Ustundag, Bulent Guloglu (2017), “A Forecasting Approach for Truckload Spot Marketing Pricing”, studies that Logistic and supply chain has now become one of the largest growing and competitive sector with third party associations and outsourcing in the modern world. Therefore a sound and dependable forecasting technique is needed by the companies and Logistic service providers for pricing truckload spot market in supply chain.
Through this research a framework for a price forecasting model for truckload spot market is developed and applied. In development of framework artificial neural network and quantile regression are used and this framework is tested in route based model specifically and also in general model where all routes can be considered together.
Based on the result he best approach can be selected to forecast the trucking prices.
9. Hesam Dehghania, Dejan Bogdanovic (2017), “Copper Price Estimation Using Bat Algorithm” Then, the estimation parameters of the equation was modified using Bat algorithm. Finally, it is concluded that the determined equation with 0.132 of RMSE can predict the copper price better than the classic estimation methods. Mining industry and its performance is highly dependent on metal price volatility and predicting it with accuracy is critical for the mining projects. Classical estimation models are now obsolete as they are not able to correctly estimate the volatility due to ever changing market demands and variations since last few years.
Therefore there is a need to use modern estimation models and algorithms for better prediction of price volatility. So copper price estimation was done through BAT algorithm and time series function. Brownian motion with mean reversion (BMMR) was found to be the best with the root mean square error (RMSE) of 0.449 which was modified with BAT algorithm to get RMSE as 0.132 which is closer to reality as compared to classic BMMR?
10. Alessio Azztti (2016), “Forecasting Gold Price: A Comparative Study”, this research evaluates the correctness of various existing forecasting techniques and models to provide accurate gold prices forecast. It was found that ARIMA model provided the best forecast of prices over 36 month forecasting horizons as compared to other models. ARIMA model provided better result to metal such as silver, platinum, palladium and rhodium as well.
11. Banhi Guha and Gautam Bandyopadhyay (2016), “Gold Price Forecasting Using ARIMA Model”, studied that gold is widely used over a long period of time by every class of people for investment purpose in beneficial outcomes. So its essential forecast the gold prices over a period of time. In this study ARIMA time series application model is used to forecast
the gold prices based on past data from November 2003 to January 2014 to evaluate the risk in purchase of gold which help investors to know the suitable time for buy and selling the gold. It helps in dealing with adverse economical factors.
12. Asad Ali1, Muhammad Iqbal Ch., Sadia Qamar, Noureen Akhtar, Tahir Mahmood, Mehvish Hyder, Muhammad Tariq Jamshed (2016), “Forecasting of Daily Gold Price By Using Box-Jenkins Methodology”, In this research, time series model used for forecasting the daily gold price and use set of data of united state dollars per ounce from Jan 02, 2014 to July 03, 2015 for in this research. With the help of Box- Jenkins methodology, ARIMA model is selected and the model selection criterion shows that ARIMA (1,1,0) and ARIMA (0,1,1) are close to each other for forecasting the daily gold price.
Now, after calculation we find that ARIMA (0,1,1) is more efficient than ARIMA (1,1,0) on the base model selection criteria’s MAE, MAPE, RMSE.
13. N.A.M.R. Senaviratna (2016), “Forecasting Tea Auction Prices In Sri Lanka: Box-Jenkins Modelling Approach”, studied that, it is essential to forecast the auction price and also develop the acceptable model to forecast the tea auction price in Sri Lanka because Sri Lanka is the 3rd biggest exporter and 4th biggest manufacturer of tea in the world. In this research, Box- Jenkins model use to forecast the auction tea price in Sri Lanka over a period from 1996 to 2016. According to AIC, SBC, MSE and R2 the most acceptable model is Seasonal Autoregressive Integrated Moving Average (SARIMA) (1,0,0) (0,1,0)12 to forecast the tea auction price in Sri Lanka. The ADF test, LM test, White's General test and Jarque-Bera test was used to check the acceptability of fitted model and also mean absolute percentage error (MAPE) criteria used to measure the forecasting performance of the fitted model.
14. Dr. Muhammad Iqbal Ch., Muhammad Tariq Jamshaid, Asad Ali and Qaisar Rashid (2016), “Forecasting of Wheat Production: A Comparative Study of Pakistan And India”, the aim of this research to forecasting the annual wheat production of India and Pakistan with the help of secondary data from 1960 to 2015 in 100MT. Box- Jenkins method &
ARIMA model was used to forecast the wheat production of India and Pakistan. The model selection criterion (AIC & SBC) indicates that ARIMA (1,1,1) & (0,1,1) providing good performance for Pakistan where ARIMA (1,1,0) & (0,1,1) are better for India’s annual wheat production forecast. The result of this research shows that ARIMA (1,1,1) & (1,1,0) are more accurate for India and Pakistan.
15. Mohamed M. Mostafa, Ahmed A. El-Masry (2015), “Oil Price Forecasting Using Gene Expression Programming And Artificial Neutral Networks”, the aim of this research is to forecast the oil prices with the help of Gene Expression Programming (GEP) and Artificial Neutral Network (NN) models from ranging Jan 2, 1986 to June 12, 2012 is used. In this research, ARIMA model are used as a benchmark models and the result shows that GEP model also performs the NN &
ARIMA models in form of mean square error, root mean square error and mean absolute error. R- Square statistic is used to measured the highest explanatory power of GEP model.
16. M. Khalid, Mariam Sultana, Faheem Zaidi (2014), “Forecasting Gold Price: Evidence From Pakistan Market” In this research, gold price are forecasted with the use of ARIMA and two version of wavelet scheme. The monthly data is studies consisting of 221 observations from Dec 2005 to April 2013. After comparison it is found that wavelet neural transformation is more accurate than other models by using mean absolute error and mean square error.
17. Vinayak N. Jalikatti, Raghvendra Chourad, Najeer Ahmad D.G., Shreya Amarapurkar and Sarfaraz Sheikh (2014),
“Forecasting The Prices of Onion In Belgaum Market of Northern Karnataka Using ARIMA Technique”, the aim of this research is to forecasting the onion price at Belgaum district of Northern Karnataka by using ARIMA model or technique.
The result of this research indicate the ex- ante & ex- post forecast of monthly onion price for Belgaum market. In this market, it was noticed that there was rapid increase in the prices during 1999, 2000 & 2011 and also year- wise decrease in production and sufficient storage facilities might be the reasons for rapid increase in the onion price. These forecasted prices help the farmers to plan the production process in advance.
18. Christian Pierdzioch, Jan-Christoph Rulke, and Georg Stadtmann (2012), “Forecasting Metal Prices: Do Forecasters Herd?” Said that metals are crucial and major source of export revenue for developing countries and changes in metal prices have a impact on trade of country. So it is essential to track the metal prices. Analysis of 20,000 Forecasts of nine metal prices are carried out at four different horizons. As a result it is found that forecasts are heterogeneous due to anti- herding and it scatters forecast around a general agreement.
19. Shahriar shafiee Erkan Tapal (2010), “An Overview of Global Gold Market And Gold Price Forecasting"studied that it is essential to estimate future prices of metal to minimize the risk on global economy at the time of adverse situation such as recession with the use of various forecasting models. In this research we forecast future price trends of gold in global market.
The first section studies the historical trend of gold prices from Jan 1968 to Dec 2008. Then an investigation is done on investigation between gold prices and oil price and global inflation over last 40 years. The second section deals with jump and dip diffusion model for forecasting commodity prices of natural resources. This study validates model and estimate gold
20. Viviana Fernandez, “Forecasting Commodity Prices By Classification Methods: The Cases of Crude Oil And Natural Gas Spot Prices” studies that forecasting economic activities is essential since long term in the area of natural resources. This project help in forecasting commodity price that is crude oil and natural gas on daily basis. The techniques used for forecasting prices are ANN and SVM. They use ARIMA as a bench work for forecasting in correspondence to ANN & SVM specification for 2-4 days and 10-20 days for more accurate forecasting than individual forecasting. By this method we reach on accurate conclusion of commodity price forecasting.
III. PROPOSEDWORK
The Gold is the very precious and costly metal. So, we forecast the gold price in advance which helps the people to purchase the gold at suitable time. For this purpose, we use different modern forecasting techniques for forecast the gold prices and also compare the different forecasting techniques for finding the best technique or method to forecast the gold price.
IV. CONCLUSION
The above studied shows that the number of researchers used different forecasting techniques for forecast the oil price, metal price, agricultural product price etc. But most of the researchers use ARIMA model for forecasting the different metals prices and also compare the various methods of forecasting. The gold is the very important metal, so various researchers use ARIMA model and Box-Jenkins model for forecasting the gold price and also, this model provide best gold price forecasting.
V. ACKNOWLEDGEMENT
It gives me immense pleasure to express my deepest sense of gratitude and sincere thanks to my highly respected and esteemed guide Mr. Tarun Kumar Yadav, MECHANICAL ENGINEERING DEPARTMENT, BTIRT, Sagar, for their valuable guidance, encouragement and help for completing this work. Their useful suggestions for this whole work and co- operative behavior are sincerely acknowledged.
I would like to express my sincere thank to Mr. Akash Tomar, MECHANICAL ENGINEERING DEPARTMENT, BTIRT, Sagar for giving me this opportunity to undertake this Research.
I also wish to express my indebtedness to my parents as well as my family member whose blessings and support always helped me to face the challenges ahead.
REFERENCES
1. Syazwani Zainal Abidin et al. (2018), “Comparison on Estimating Malaysia Gold Price via Nonlinear Prediction Method and Box–Jenkins Model”,
2. Hesam Dehghani and Mahsa Zangeneh (23 July 2018), “Crude Oil Price Forecasting: A Biogeography-Based Optimization Approach”, ISSN: 1556-7249 (Print) 1556-7257 (Online)
3. Naliniprava Tripathy (2017), “Forecasting Gold Price with Auto Regressive Integrated Moving Average Model”, ISSN: 2146- 4138, IJEFI, 2017, 7(4), 324-329.
4. Sneha Chaudhry et al. (June 2017) “A Time Series Analysis of Auction Prices of Indian Tea”, (IJRESS) Vol. 7 Issue 6, June- 2017, pp. 100~111 ISSN(o): 2249-7382 | Impact Factor: 6.93
5. A Rahman and A S Ahmar (Sept.2017), “Forecasting of Primary Energy Consumption Data in the United States: A Comparison Between ARIMA and Holter-Winters Models”, AIP Conference Proceedings 1885, 020163 (2017)
6. Paulo V. Cenas (Nov 2017), “Forecast of Agricultural Crop Price using Time Series and Kalman Filter Method”, Asia Pacific Journal of Multidisciplinary Research Vol. 5 No.4, 15-21 November 2017 Part III P-ISSN 2350-7756 E-ISSN 2350-844 7. Syed Misbah Uddin et al. (Dec 2017), “Comparison of Some Statistical Forecasting Techniques with GMDH Predictor: A
Case Study”, Journal of Mechanical Engineering, Vol. ME 47
8. Aysenur Budak et al. (2017), “A Forecasting Approach for Truckload Spot Marketing Pricing”, Elsevier, vol. 97(c), pages 55- 68
9. Hesam Dehghania, Dejan Bogdanovic (2017), “Copper Price Estimation Using Bat Algorithm”, Resources Policy xxx (xxxx) xxx-xxx
10. Alessio Azztti (27 Feb 2016), “Forecasting Gold Price: A Comparative Study”
11. Banhi Guha and Gautam Bandyopadhyay (March 2016), “Gold Price Forecasting Using ARIMA Model”, Journal of Advanced Management Science Vol. 4, No. 2, pp. 117–121
12. Asad Ali1 et al. (2016), “Forecasting of Daily Gold Price By Using Box-Jenkins Methodology”, International Journal of Asian Social Science ISSN(e): 2224-4441/ISSN(p): 2226-5139
13. N.A.M.R. Senaviratna (2016), “Forecasting Tea Auction prices in Sri Lanka: Box-Jenkins Modeling Approach”, International Journal of Applied Research, Volume : 6 | Issue : 11, November 2016, ISSN - 2249-555X, IF : 3.919, IC Value : 74.50 14. Dr. Muhammad Iqbal Ch. et al. (2016), “Forecasting of Wheat Production: A Comparative Study of Pakistan & India”,
International Journal of Applied Research 4(12), 698-709, ISSN: 2320-5407
15. Mohamed M. Mostafa, Ahmed A. El-Masry (2015), “Oil Price Forecasting Using Gene Expression Programming And Artificial Neutral Networks”,
16. M. Khalid, Mariam Sultana, Faheem Zaidi (2014), “Forecasting Gold Price: Evidence From Pakistan Market”, ISSN 2222- 1697 (Paper) ISSN 2222-2847 (Online) Vol.5, No.3
17. Vinayak N. Jalikatti et al. (2014), “Forecasting The Prices of Onion in Belgaum Market of Northern Karnataka Using ARIMA Technique”, International Research Journal of Agricultural Economics & Statistics, Volume 5, Issue 2, September, 2014 , 153-159 , e ISSN-2231-6434
18. Christian Pierdzioch et al. (2012), “Forecasting Metal Prices: Do Forecasters Herd?”, ISSN 1860 0921, Paper No. 325
19. Shafiee S. and Topal E. (2010), “An Overview of Global Gold Market and Gold Price Forecasting", Resources Policy 35(3), pp. 178-189.
20. Viviana Fernandez , “Forecasting Commodity Prices By Classification Methods: The Cases of Crude Oil And Natural Gas Spot Prices”
21. https://brainly.in/question/1618623, Time: 8.15 pm, Date: 20-11-2018
