Seasonal Autoregressive Integrated Moving Average Model Research Articles

Long-term monthly average temperature forecasting in some climate types of Iran, using the models SARIMA, SVR, and SVR-FA

Temporal changes of the global surface temperature have been used as a prominent indicator of global climate change; therefore, making dependable forecasts underlies the foundation of sound environmental policies. In this research, the accuracy of the Seasonal Autoregressive Integrated Moving Average (SARIMA) Stochastic model has been compared with the Support Vector Regression (SVR) and its merged type with Firefly optimization algorithm (SVR-FA) as a meta-innovative model, in long-term forecasting of average monthly temperature. For this, 5 stations from different climates of Iran (according to the Extended De Martonne method) were selected, including Abadan, Anzali, Isfahan, Mashhad, and Tabriz. The data were collected during 1951–2011, for training (75%) and testing (25%). After selecting the best models, the average monthly temperature has been forecasted for the period 2012–2017. The results showed that the models had better performances in Extra-Arid and Warm (Abadan) and after that Extra-Arid and Cold (Isfahan) climate, in long-term forecasting. The weakest performances of the models were reported in Semi-Arid and Cold climate, including Mashhad and Tabriz. Also, despite the use of the non-linear SVR model and its meta-innovative type, SVR-FA, the results showed that, in the climates of Iran, the linear and classical SARIMA model still offers a more appropriate performance in temperature long-term forecasting. So that it could forecast the average monthly temperature of Abadan with root mean square error (RMSE) = 1.027 °C, and Isfahan with RMSE = 1.197 °C for the 6 years ahead. The SVR and SVR-FA models also had good performances. The results of this checking also report the effectiveness of the merging SVR model with the Firefly optimization algorithm in temperature forecasting in Iran’s climates, so, compared with the SVR model, it is suggested to use SVR-FA for temperature forecasting.

Time series analysis of human brucellosis in mainland China by using Elman and Jordan recurrent neural networks

BackgroundEstablishing epidemiological models and conducting predictions seems to be useful for the prevention and control of human brucellosis. Autoregressive integrated moving average (ARIMA) models can capture the long-term trends and the periodic variations in time series. However, these models cannot handle the nonlinear trends correctly. Recurrent neural networks can address problems that involve nonlinear time series data. In this study, we intended to build prediction models for human brucellosis in mainland China with Elman and Jordan neural networks. The fitting and forecasting accuracy of the neural networks were compared with a traditional seasonal ARIMA model.MethodsThe reported human brucellosis cases were obtained from the website of the National Health and Family Planning Commission of China. The human brucellosis cases from January 2004 to December 2017 were assembled as monthly counts. The training set observed from January 2004 to December 2016 was used to build the seasonal ARIMA model, Elman and Jordan neural networks. The test set from January 2017 to December 2017 was used to test the forecast results. The root mean squared error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) were used to assess the fitting and forecasting accuracy of the three models.ResultsThere were 52,868 cases of human brucellosis in Mainland China from January 2004 to December 2017. We observed a long-term upward trend and seasonal variance in the original time series. In the training set, the RMSE and MAE of Elman and Jordan neural networks were lower than those in the ARIMA model, whereas the MAPE of Elman and Jordan neural networks was slightly higher than that in the ARIMA model. In the test set, the RMSE, MAE and MAPE of Elman and Jordan neural networks were far lower than those in the ARIMA model.ConclusionsThe Elman and Jordan recurrent neural networks achieved much higher forecasting accuracy. These models are more suitable for forecasting nonlinear time series data, such as human brucellosis than the traditional ARIMA model.

Estimates of reference evapotranspiration in the municipality of Ariquemes (RO) using neural networks GMDH-type

ABSTRACT Reference evapotranspiration is a climatological variable of great importance for water use dimensioning in irrigation methods. In order to contribute to the climatic understanding of Ariquemes, Rodônia state, Brazil, the study aims to model the behavior of the time series of reference evapotranspiration using a GMDH-type (Group Method of Data Handling) artificial neural network (ANN) and to compare it with the SARIMA (Seasonal Autoregressive Integrated Moving Average) methodology. Data from the National Institute of Meteorology - INMET, obtained at the Automatic Weather Station of Ariquemes, from January 2011 to January 2014, were used. Data analysis was performed using software R version 3.3.1 through the GMDH-type ANN package. Modeling by GMDH-type ANN led to results similar to the results of the SARIMA model, thus constituting an option to predict climatic time series. GMDH-type models with larger numbers of inputs and layers presented lowest mean square error.

Exploring the trend, prediction and driving forces of aerosols using satellite and ground data, and implications for climate change mitigation

Human activities-related aerosol emissions and CO2 emissions originate from many of the common sources. Identifying the aerosol variations and the underling determinates can provide insights into united mitigation policy controls targeting on both aerosol pollution and climate change. Long-term trend analysis and modeling offers an effective way to fully appreciate how aerosols interlink with carbon cycle and climate change. This study analyzes the current trends, models the future predictions, and investigates potential driving forces of aerosol loading at six sites across North America and East Asia during 2003–2015. Satellite-retrieved MODIS Collection 6 retrievals and ground measurements derived from AERONET are used. Results show that there is a persistent decreasing trend in AOD for both MODIS data and AERONET data at three sites. Monthly and seasonal AOD variations reveal consistent aerosol patterns at sites along mid-latitudes. Regional differences caused by impacts of climatology and land cover types are observed for the selected sites. Statistical validation of time series ARIMA models indicates that the non-seasonal ARIMA model performs better for AERONET AOD data than for MODIS AOD data at most sites, suggesting the method works better for data with higher quality. The seasonal ARIMA model reproduces time series with distinct seasonal variations much more precisely. The reasonably predicted AOD values could provide reliable estimates to better inform the decision-making for sustainable environmental management. Drawn from aerosol pollution control strategies, it is suggested that the enforcement of regulations on emission sources and the initiative of reforestation on emission sinks could have potential implications for climate change mitigation.

Predicting the GFCF of the Brazilian construction industry: a comparison between Holt Winters' and SARIMA models

The present study focuses on creating a forecasting model in order to predict the behavior of the economic indicator known as Gross Fixed Capital Formation (GFCF) of the Brazilian construction industry, which reflects the amount of investment in the construction industry sector. The data set consists of monthly observations from January 1996 to December 2016, the year of 2016 is used as validation for the forecast model. As strong seasonality was identified in the time series, Seasonal Autoregressive Integrated Moving Average (SARIMA) and Holt Winters' models are applied and compared. After the evaluation of the selected models, the ARIMA (2,1,2) × (0,1,1) 12 is identified as the best forecast model with reasonable deviations. However, the damped multiplicative Holt Winters' model also produces good results, despite its inability in eliminating the autocorrelation in the residuals. Therefore, both models can predict the GFCF with good accuracy, which can be useful for decision-making by investors and business managers.

Parsimonious Short-Term Load Forecasting for Optimal Operation Planning of Electrical Distribution Systems

The optimal operation planning (OOP) of electrical distribution systems (EDS) is very sensible to the quality of the short-term load forecasts. Assuming aggregated demands in EDS as univariate non-stationary seasonal time series, and based on historical measurements gathered by smart meters, this paper presents a parsimonious short-term load forecasting method to estimate the expected outcomes of future demands, and the standard deviations of forecast errors. The chosen short-term load forecasting method is an adaptation of the multiplicative autoregressive integrated moving average (ARIMA) models. Seasonal ARIMA models are parsimonious forecasting techniques because they require very few parameters and low computational resources to provide an adequate representation of stochastic time series. Two approaches are used in this paper to estimate the parameters that constitute the proposed multiplicative ARIMA model: a frequentist and a Bayesian approach. Advantages and disadvantages of both methods are compared by simulating a centralized self-healing scheme of a real EDS that uses the forecasts to deploy a robust restoration plan. Results show that the proposed seasonal ARIMA model is a fast, precise, straightforward, and adaptable load forecasting method, suitable for OOP of highly supervised EDS.

Density tourism demand forecasting revisited

This study used scoring rules to evaluate density forecasts generated by different time-series models. Based on quarterly tourist arrivals to Hong Kong from ten source markets, the empirical results suggest that density forecasts perform better than point forecasts. The seasonal autoregressive integrated moving average (SARIMA) model was found to perform best among the competing models. The innovation state space models for exponential smoothing and the structural time-series models were significantly outperformed by the SARIMA model. Bootstrapping improved the density forecasts, but only over short time horizons.This article also launches the Annals of Tourism Research Curated Collection on Tourism Demand Forecasting, a special selection of research in this field.

The Use of Large-Scale Climate Indices in Monthly Reservoir Inflow Forecasting and Its Application on Time Series and Artificial Intelligence Models

Climate variability is strongly influencing hydrological processes under complex weather conditions, and it should be considered to forecast reservoir inflow for efficient dam operation strategies. Large-scale climate indices can provide potential information about climate variability, as they usually have a direct or indirect correlation with hydrologic variables. This study aims to use large-scale climate indices in monthly reservoir inflow forecasting for considering climate variability. For this purpose, time series and artificial intelligence models, such as Seasonal AutoRegressive Integrated Moving Average (SARIMA), SARIMA with eXogenous variables (SARIMAX), Artificial Neural Network (ANN), Adaptive Neural-based Fuzzy Inference System (ANFIS), and Random Forest (RF) models were employed with two types of input variables, autoregressive variables (AR-) and a combination of autoregressive and exogenous variables (ARX-). Several statistical methods, including ensemble empirical mode decomposition (EEMD), were used to select the lagged climate indices. Finally, monthly reservoir inflow was forecasted by SARIMA, SARIMAX, AR-ANN, ARX-ANN, AR-ANFIS, ARX-ANFIS, AR-RF, and ARX-RF models. As a result, the use of climate indices in artificial intelligence models showed a potential to improve the model performance, and the ARX-ANN and AR-RF models generally showed the best performance among the employed models.

Modeling of GRACE-Derived Groundwater Information in the Colorado River Basin

Groundwater depletion has been one of the major challenges in recent years. Analysis of groundwater levels can be beneficial for groundwater management. The National Aeronautics and Space Administration’s twin satellite, Gravity Recovery and Climate Experiment (GRACE), serves in monitoring terrestrial water storage. Increasing freshwater demand amidst recent drought (2000–2014) posed a significant groundwater level decline within the Colorado River Basin (CRB). In the current study, a non-parametric technique was utilized to analyze historical groundwater variability. Additionally, a stochastic Autoregressive Integrated Moving Average (ARIMA) model was developed and tested to forecast the GRACE-derived groundwater anomalies within the CRB. The ARIMA model was trained with the GRACE data from January 2003 to December of 2013 and validated with GRACE data from January 2014 to December of 2016. Groundwater anomaly from January 2017 to December of 2019 was forecasted with the tested model. Autocorrelation and partial autocorrelation plots were drawn to identify and construct the seasonal ARIMA models. ARIMA order for each grid was evaluated based on Akaike’s and Bayesian information criterion. The error analysis showed the reasonable numerical accuracy of selected seasonal ARIMA models. The proposed models can be used to forecast groundwater variability for sustainable groundwater planning and management.

Seasonality and Trend Forecasting of Tuberculosis Incidence in Chongqing, China.

Tuberculosis (TB) is a global infectious disease and one of the ten leading causes of death worldwide. As TB incidence is seasonal, a reliable forecasting model that incorporates both seasonal and trend effects would be useful to improve the prevention and control of TB. In this study, the X12 autoregressive integrated moving average (X12-ARIMA) model was constructed by dividing the sequence into season term and trend term to forecast the two terms, respectively. Data regarding the TB report rate from January 2004 to December 2015 were included in the model, and the TB report data from January 2016 to December 2016 were used to validate the results. The X12-ARIMA model was compared with the seasonal autoregressive integrated moving average (SARIMA) model. A total of 383,797 cases were reported from January 2004 to December 2016 in Chongqing, China. The report rate of TB was highest in 2005 (151.06 per 100,000 population) and lowest in 2016 (72.58 per 100,000 population). The final X12-ARIMA model included the ARIMA (3,1,3) model for the trend term and the ARIMA (2,1,3) model for the season term. The SARIMA (1,0,2) * (1,1,1)12 model was selected for the SARIMA model. The mean absolute error (MAE) and mean absolute percentage error (MAPE) of fitting and predicting performance based on the X12-ARIMA model were less than the SARIMA model. In conclusion, the occurrence of TB in Chongqing is controlled, which may be attributed to socioeconomic developments and improved TB prevention and control services. Applying the X12-ARIMA model is an effective method to forecast and analyze the trend and seasonality of TB.

Modelling Electricity Demand in Sudan Using Multiplicative Seasonal ARIMA and HOLT-WINTERS

Reliable forecast of energy demand represents a starting point in policy development and improvement of production and distribution facilities. This paper predicts the electricity demand in Sudan during the period from January 2006 to December 2016. For the purposes of these forecasts, multiplicative seasonal ARIMA and Holt-Winters models were used. The results show that the Holt-Winters exponential smoothing method, as opposed to one. An autoregressive model has been included in both the standard and the seasonal Holt-Winters methods to model the residuals and thus improve the forecast accuracy.

Application of a combined model with seasonal autoregressive integrated moving average and support vector regression in forecasting hand-foot-mouth disease incidence in Wuhan, China.

Hand-foot-mouth disease (HFMD) is a serious public health problem with increasing cases and substantial financial burden in China, especially in Wuhan city. Hence, there is an urgent need to construct a model to predict the incidence of HFMD that could make the prevention and control of this disease more effective.The incidence data of HFMD of Wuhan city from January 2009 to December 2016 were used to fit a combined model with seasonal autoregressive integrated moving average (SARIMA) model and support vector regression (SVR) model. Then, the SARIMA-SVR hybrid model was constructed. Subsequently, the fitted SARIMA-SVR hybrid model was applied to obtain the fitted HFMD incidence from 2009 to 2016. Finally, the fitted SARIMA-SVR hybrid model was used to forecast the incidence of HFMD of the year 2017. To assess the validity of the model, the mean square error (MSE) and mean absolute percentage error (MAPE) between the actual values and predicted values of HFMD incidence (2017) were calculated.From 2009 to 2017, a total of 107636 HFMD cases were reported in Wuhan City, Hubei Province, and the male-to-female ratio is 1.60:1. The age group of 0 to 5 years old accounts for 95.06% of all reported cases and scattered children made up the large proportion (accounted for 56.65%). There were 2 epidemic peaks, from April to July and September to December, respectively, with an emphasis on the former. High-prevalence areas mainly emerge in Dongxihu District, Jiangxia District, and Hongshan District. SARIMA (1,0,1)(0,0,2)[12] is the optimal model given with a minimum Akaike information criterion (AIC) (700.71), then SVR model was constructed by using the optimum parameter (C = 100000, =0.00001, =0.01). The forecasted incidences of single SARIMA model and SARIMA-SVR hybrid model from January to December 2017 match the actual data well. The single SARIMA model shows poor performance with large MSE and MAPE values in comparison to SARIMA-SVR hybrid model.The SARIMA-SVR hybrid model in this study showed that accurate forecasting of the HFMD incidence is possible. It is a potential decision supportive tool for controlling HFMD in Wuhan, China.

Generating electrical demand time series applying SRA technique to complement NAR and sARIMA models

Prediction of demand time series is commonly approached to deliver a punctual forecast model, however, is highly recommended to provide probabilistic models that give a range to each future value of forecasting horizon. In this paper, three demand series are analyzed and forecasted by a non-linear autoregressive (NAR24) neural network and a seasonal ARIMA (sARIMA) model from which naive prediction intervals (PI) are also computed. The error measurement for the forecast models indicated adequate accuracy and forecast performance (MAPE values under 4%), in general. As the major innovation to the literature, a methodology to provide complementary limit stochastic scenarios (LSS) for each forecast model is presented with two approaches: high consumption (HCA) and low consumption (LCA). Using fractal Brownian motion (fBm) concepts and the successive random addition technique (SRA), random walks (RW) were simulated and then added to both forecast models to generate stochastic scenarios. To start a random walk, three input parameters were determined for each case study: range, length, and Hurst coefficient (H). The most probable stochastic scenarios (PSS) accordingly to the variation coefficient were selected from all the produced scenarios. The PSS with the highest and the lowest average were selected as the limit stochastic scenarios, LSSmin and LSSmax, respectively. Total energy for the following 24 h was calculated and it showed that the range provided by LSS delivers additional information to electricity dispatchers accordingly to HCA and LCA situations which nor forecast models nor PI can foresee. Finally, in order to compare the LSS, the maximum and minimum limit scenarios were averaged to produce a stochastic “model” for each case study. Using common error measurements MAE, MSE, and MAPE, the LSS applied to NAR24 demonstrated to be more reliable in two out of three case studies.

Undergraduate International Student Enrollment Forecasting Model: An Application of Time Series Analysis

This study developed statistical models to forecast international undergraduate student enrollment at a Midwest university. The authors constructed a SARIMA (Seasonal Autoregressive Integrated Moving Average) model with input variables to estimate future enrollment. The SARIMA model reflected enrollment patterns by semester through highlighting seasonality. Further, authors added input variables such as visa policy changes, the rapid increase of Chinese undergraduate enrollment, and tuition rate into the model estimation. The visa policy change and the increase of Chinese undergraduate enrollment were significant predictors of international undergraduate enrollment. The effect of tuition rates was significant but minimal in magnitude. Findings of this study generate significant implications for policy, enrollment management, and student services for international students.

Forecasting of Potato Prices of Hooghly in West Bengal: Time Series Analysis Using SARIMA Model

Potato is one of the most important food crops in India. Potato is also one of the principal cash crop, it gives handsome returns to the farmers. West Bengal is a state in India where potato is one of the major agricultural crops. But potato markets are more volatile due to fluctuation in production rather than Stable. So price of potato is fluctuating in nature. Thus, the time series analysis and price forecast may help producers in acreage allocation and timing of sale of potato. The present study was conducted to know the statistical investigation of price behaviour of potato in Hooghly district of West Bengal. In the present study, Box-Jenkins Seasonal Auto Regressive Integrated Moving Average (SARIMA) modeling is deployed in forecasting of monthly average price of potato in Hooghly of West Bengal up to October 2020 based on data from November 2008 to October 2018 (a period of 120 months). Seasonal indices calculated showed that generally the price is low from January to April and it starts picking up from May and reaches the maximum in November. The best model has been selected based on the Bayesian Information Criteria (BIC), Root Means Square Error (RMSE), Mean Absolute Percent Error (MAPE), Mean Absolute Error (MAE) and highest R-Square. The estimated best SARIMA model is (1,1,0)(4,1,0)12. Root Mean Square Error (RMSE), Mean Average Percentage Error (MAPE), Mean Absolute Error (MAE) and BIC for Hooghly district are 184.10, 19.08, 118.98 and 10.69 respectively. Short term forecasts based on this model are close to the observed values and the behaviour of forecasted price of potato truly reflected the actual price as well as market tendency.

Employing seasonal autoregressive integrated moving average forecasting model to predict the number of dengue cases in Mumbai

Background: Dengue control suffers challenges such as the absence of specific treatment and lack of vaccine. Forecasting of dengue would facilitate allocation of resources that are needed for such activities. The seasonal autoregressive integrated moving average (SARIMA) is the mathematical model, which provides the estimated monthly figures for the given period. Objectives: The objectives of the study were to select the best prediction model for dengue fever (DF) by time series data over the past 13 years in the Mumbai city and to forecast monthly dengue incidence for 2019. Materials and Methods: Retrospective study design was employed at epidemiology cell, Mumbai – Integrated Disease Surveillance Project (IDSP). The reported DF/dengue hemorrhagic fever cases during January 2006–December 2018 were mobilized from epidemiology cell of the city. Data were recorded on Excel sheet. The SARIMA model was applied to the data with R software. Results: The cases showed a form with a seasonal difference. SARIMA (1, 1, 2)(0, 1, 1) model had the highest Akaike information criteria (AIC) of 1433.18 and mean absolute percentage error of 43.75 and performed to be the best model. Adequacy of the model was established through the Ljung–Box test, which showed no substantial correlation among residuals at different lag times. Seasonal peak is estimated in the month of September of 2019 with 323 (95% confidence interval: [196.56, 449.92]) cases followed by 222 (95% confidence interval: [95.58, 348.93]) in October. Conclusions: The function of the SARIMA model may be useful for a forecast of cases and impending outbreaks of dengue.

Rainfall and runoff estimation of micro watersheds of coastal Navsari

The onset, withdrawal and quantity of rainfall greatly influence the agricultural yield, economy, water resources, power generation and ecosystem. Hence, if the variations in rainfall are known well in advance, it would be possible to reduce the adverse impacts related to excess or deficient rainfall, providing us prior information about droughts and floods. The principles of stochastic processes have been increasingly and successfully applied in the past three decades to model many of the hydrological processes which are stochastic in nature. Time series modelling has been extensively used in stochastic hydrology for predicting hydrological processes like rainfall and runoff. Time lagged models extract maximum possible information from the available record for forecasting. In this study, rainfall forecast of Navsari was attempted using rainfall training data of 30 years (1983-2012) and runoff was estimated using curve number method for each micro watershed of coastal Navsari. The models used for forecasting rainfall, based on the time series data, were seasonal ARIMA model and ANN multilayer perceptron model. A seasonal ARIMA model, based on Box Jenkins method model was built by determining its appropriate parameters and the ANN model was built using Levenberg Marquardt algorithm. The seasonal ARIMA model was determined using four steps namely identification of autoregressive and moving average terms, estimating the parameters using maximum likelihood method, diagnostic checking of residuals and forecasting. A number of ANN model structures were tested and the appropriate ANN model was selected based on its performance on validation data after which the model was deployed for forecasting. It was concluded that the non linear ANN model was superior to linear ARIMA model in terms of its forecasting accuracy tested on validation data. The predicted rainfall was then utilized to estimate the runoff using Curve number method for each micro watershed of Navsari coast.

Comparison of Models for Excess Mortality of Influenza Applied to Japan

To elucidate the social effects of an influenza outbreak, the World Health Organization recommends a concept for excess mortality attributable to an influenza outbreak. However, because several models exist to estimate excess mortality, we would like to ascertain the most appropriate of three models: the Center for Disease Control and Prevention (CDC) model, the seasonal autoregressive integrated moving average (SARIMA) model, and the National Institution of Infectious Diseases (NIID) model. Excess mortality is defined as the difference between the actual number of deaths and the epidemiological threshold. The epidemiological threshold is defined as upper bound of 95% confidence interval (CI) of the baseline. The actual number of deaths might be less than the baseline, which implies inconsistent with the definition of baseline. Especially, actual deaths fewer than the lower bound of 95% CI of baseline suggest the inappropriateness of a model of excess mortality. Among 123 months during epidemic periods, the NIID model found excess mortality in 56 months, CDC model in 31 months, and SARIMA model in 35 months. Conversely, the NIID model found negative excess mortality in only 2 months, but the CDC model and SARIMA model found it respectively for 10 and 33 months. Negative excess mortality represents the logical inconsistency of the model. Therefore, NIID model might be the best among the three models considered.

Predicting number of traumas using the seasonal time series model

Background: Road accidents and casualties resulted are among the current challenges of human societies, which have imposed a high cost on the economies of countries. Objectives: Prediction of accidents caused by driving incidents helps planners achieve a suitable model to reduce the occurrence of traumas resulted from the driving accidents. Materials and Methods: In this study, a seasonal time series model was used for predicting the number of road accident traumas. Data related to the patients referring to Imam Khomeini Hospital in Ilam Province were evaluated from March 2012 to June 2017. Results: The results showed that during November and October in 2015 and 2016, we had the highest number of accidents due to high traffic during New Year's Vacation, summer trips, and religious pilgrimages including the Arbaeen. Moreover, the results depicted that the seasonal Arima model was effective in predicting the number of traumas due to accidents. Furthermore, forecasting the model showed an ascending trend in the number of accidents in the following 3 years. Conclusion: The number of accident traumas in the forthcoming months can be predicted through time series models. Of course, these models can be used by managers as appropriate tools for traffic analysis. Furthermore, the increasing trend in the number of traumas indicates that serious consideration for planning and managing such events seems necessary for the administrators in the near future.

Forecasting residential building costs in New Zealand using a univariate approach

Construction cost index has been widely used to prepare cost estimates, budgets, and bids for construction projects. It can also be regarded as an indicator of cost level, which makes it valuable to public authorities for understanding the conditions in the construction industry. Accurate forecasting of future construction cost index is essential for construction industry at both micro- and macro-level. To improve the accuracy of the cost forecasting, time series modeling techniques are adopted in this study. The performance of the exponential smoothing models and seasonal autoregressive integrated moving average (ARIMA) models for forecasting the building cost of five categories of residential building (one-story house, two-story house, town house, apartment, and retirement village building) in New Zealand is compared. Exponential smoothing models can produce more accurate forecasts for cost series of the one-story house and two-story house in New Zealand, while seasonal ARIMA models outperform exponential smoothing models across the cost series for town house, apartment, and retirement village building. This study contributes toward the development of the current state of knowledge in the area of cost index forecasting for New Zealand and provides insights that should be valuable from the practitioner perspectives.