Mariano Test Research Articles

Climate-Based Modeling and Prediction of Rice Gall Midge Populations Using Count Time Series and Machine Learning Approaches

The Asian rice gall midge (Orseolia oryzae (Wood-Mason)) is a major insect pest in rice cultivation. Therefore, development of a reliable system for the timely prediction of this insect would be a valuable tool in pest management. In this study, occurring between the period from 2013–2018: (i) gall midge populations were recorded using a light trap with an incandescent bulb, and (ii) climatological parameters (air temperature, air relative humidity, rainfall and insulations) were measured at four intensive rice cropping agroecosystems that are endemic for gall midge incidence in India. In addition, weekly cumulative trapped gall midge populations and weekly averages of climatological data were subjected to count time series (Integer-valued Generalized Autoregressive Conditional Heteroscedastic—INGARCH) and machine learning (Artificial Neural Network—ANN, and Support Vector Regression—SVR) models. The empirical results revealed that the ANN with exogenous variable (ANNX) model outperformed INGRACH with exogenous variable (INGRCHX) and SVR with exogenous variable (SVRX) models in the prediction of gall midge populations in both training and testing data sets. Moreover, the Diebold–Mariano (DM) test confirmed the significant superiority of the ANNX model over INGARCHX and SVRX models in modeling and predicting rice gall midge populations. Utilizing the presented efficient early warning system based on a robust statistical model to predict the build-up of gall midge population could greatly contribute to the design and implementation of both proactive and more sustainable site-specific pest management strategies to avoid significant rice yield losses.

Sticky Stock Market Analysts

Technological progress in recent years has made new methods available for making forecasts in a variety of areas. We examine the success of ex-ante stock market forecasts of three major stock market indices, i.e., the German Stock Market Index (DAX), the Dow Jones Industrial Index (DJI), and the Euro Stoxx 50 (SX5E). We test whether the forecasts prove true when they reach their effective dates and are therefore suitable for active investment strategies. We revive the thoughts of the American sociologist William Fielding Ogburn, who argues that forecasters consistently underestimate the variability of the future. In addition, we draw on some contemporary measures of forecast quality (prediction-realization diagram, test of unbiasedness, and Diebold–Mariano test). We reveal that (a) unusual events are underrepresented in the forecasts, (b) the dispersion of the forecasts lags behind that of the actual events, (c) the slope of the regression lines in the prediction-realization diagram is <1, (d) the forecasts are highly biased, and (e) the quality of the forecasts is not significantly better than that of naïve forecasts. The overall behavior of the forecasters can be described as “sticky” because their forecasts adhere too strongly to long-term trends in the indices and are thus characterized by conservatism.

Neighborhood rough set-based multi-attribute prediction approach and its application of gout patients

Accurate disease prediction is an effective way to reduce medical costs. Due to the difference of eating habits and physical fitness of patients, the traditional disease prediction methods are facing an enormous challenge. How to find a reliable disease prediction method in the uncertain environment and improve the accuracy of prediction will be a valuable scientific problem. To obtain accurate prediction and help patients reduce medical costs, this paper introduces neighborhood rough set into multivariate variational mode decomposition, and proposes a new multi-attribute prediction approach. Firstly, to avoid the interference of redundant attributes, a multi-attribute reduction method based on neighborhood rough set is established. Then, to reduce the volatility and complexity of multi-attribute data in hybrid information system, a neighborhood rough set-based multivariable variational mode decomposition method is constructed. Subsequently, a predictor of extreme learning machine with kernel function, clearly defining the mapping relationship, is developed. Furthermore, Diebold–Mariano (DM) test and probability density distribution are used to evaluate the prediction results. Finally, 2041 random physical examination samples of potential gout patients are utilized to verify the effectiveness and practicability of the proposed approach. Experimental results show that the neighborhood rough set-based multi-attribute prediction approach has high accuracy and stability. Meanwhile, a new quantitative theory and method for chronic disease management decision-making can be provided in medical decision-making.

The power of deterministic option-implied trees in pricing European options

ABSTRACT The aims of the current article are threefold. First, to investigate the power of deterministic option-implied trees, constructed either by forward or by backward induction, in pricing European options, in order to assess the proper representation of the smile. Second, to investigate and contrast the power of deterministic option-implied trees during tranquil and volatile market conditions. Last, to assess the correctness of the representation of the smile in different parts of the risk-neutral distribution. Three main results are obtained. First, the pricing performance of the Enhanced Derman and Kani model (EDK), based on forward induction, is superior to that of the Rubinstein model, based on backward induction. Second, the EDK model produces better results (smaller errors) on the left tail of the distribution, i.e. it is better in pricing out-of-the-money put options. Third, it performs better in turmoil periods where correct pricing a challenge, and accuracy is of greater importance than in tranquil periods. Diebold and Mariano test of equal predictive accuracy confirms the superiority of the EDK model in both sub-periods.

Wheat Futures Prices Prediction in China: A Hybrid Approach

Stocks markets play their financial roles of price shocks and hedging just when they are proficient. The imperative highlights of productive market are that one cannot make extraordinary profit from the stocks markets. This research investigates whether China wheat futures price can be predicted by employing artificial intelligence neural network. This would add to our knowledge whether wheat futures market is resourceful and would enable traders, sellers, and investors to improve cost-effective trading strategy. We utilize the traditional financial model to forecast the wheat futures price and acquire out of sample point estimates. We additionally assess the robustness of our outcomes by applying several alternative forecasting techniques such as artificial intelligence with one hidden layer and autoregressive integrated moving average (ARIMA) model. Furthermore, the statistical significance of our point estimation was further tested through the Mariano and Diebold test. Considering random walk forecast as the bench mark, we used a number of economic indicators, trader’s expectation towards futures prices, and lagged value of futures price of wheat in order to forecast the evaluation of wheat futures price. The computable significance of out of sample estimations recommends that our ANN with one hidden layer has the best anticipating presentation among all the models considered in this exploration and has the estimating power in foreseeing wheat futures returns. Furthermore, this investigation discovers that the futures price of wheat can be predicted, and the wheat futures market of China is not productive.

Measuring multi‐volatility states of financial markets based on multifractal clustering model

AbstractMeasuring multi‐volatility states is an important component of financial risk management. In this paper, taking 17 years' high‐frequency data of the Shanghai Stock Exchange Composite Index (SSEC) as an example, we set out to propose a novel model of multifractal clustering model (MCM), combining multifractal algorithm with k‐mean clustering algorithm, to measure the multi‐volatility states of financial markets. The empirical results present that the financial markets are multifractal and the multifractal parameter Sα measured by multifractal algorithm correlates well with the performance of financial markets. Meanwhile, the multi‐volatility states recognized by k‐mean clustering algorithm based on Sα have obvious statistical significance. More importantly, the experiment results show that by the loss functions and Diebold–Mariano (DM) test, generalized autoregressive conditional heteroscedasticity (GARCH) model with the multi‐volatility state measured by MCM is superior to others measured by Markov switching (MRS) model and hidden Markov model (HMM) as well as that without the multi‐volatility states, which indicates that MCM has best performance of measuring the multi‐volatility states of financial markets. The robustness of MCM is also evaluated by Shenzhen Securities Component Index (SZSE) instance set, SSEC instance sets with different time intervals, and different volatility models. The results still prove that MCM has strong robustness on measuring the multi‐volatility states of financial markets.

Probabilistic forecasting of photovoltaic power supply — A hybrid approach using D-vine copulas to model spatial dependencies

The fast growth of installed photovoltaic capacity is leading to an increasing impact of variable photovoltaic generation on the overall electricity industry, affecting all stakeholders in this sector. As a consequence, the importance of appropriate photovoltaic power forecasts for planning and decision support is rising, to cope with the resulting uncertainty. In particular, probabilistic forecasts are becoming increasingly important to assess the underlying risks, e.g., depicting the effect of adverse combinations. Whereas deterministic forecasts, while having the advantage of being more detailed, suffer from reflecting only an average expectation. Therefore, this paper proposes a comprehensive hybrid approach to generate deterministic and probabilistic photovoltaic power forecasts, while introducing several improvements for intra-day and day-ahead modelling and forecasting applications. In this context, several pre- and post-processing steps have been combined for the deterministic model, while the spatial interrelation of the forecasting errors is taken into account by applying D-vine copulas for the probabilistic forecasts. The reliability of the proposed hybrid approach is validated, using a comprehensive case study with high-resolution numerical weather predictions and real-world measurement data over several years for multiple photovoltaic units. Furthermore, the proposed model is benchmarked against various combinations of a photovoltaic power model (with and without statistical post-processing) and typical probabilistic models. As part of the evaluation the Energy score, Variogram-based score and Diebold–Mariano test are applied to evaluate the proposed model and highlight the strong performance of the proposed hybrid approach.

Nowcasting Growth Rates of Russia’s Export and Import by Commodity Group

In this paper, we apply a set of machine learning and econometrics models, namely: Elastic Net, Random Forest, XGBoost, and SSVS to nowcasting (estimate for the current period) the dollar volumes of Russian exports and imports by a commodity group. We use lags in the volumes of export and import commodity groups, and exchange prices for some goods and other variables, due to which the curse of dimensionality becomes quite acute. The models we use have proven themselves well in forecasting in the presence of the curse of dimensionality, when the number of model parameters exceeds the number of observations. The best-performing model appears to be the weighted machine learning model, which outperforms the ARIMA benchmark model in nowcasting the volume of both exports and imports. According to the Diebold– Mariano test, in the case of the largest commodity groups our model often manages to obtain significantly more accurate nowcasts relative to the ARIMA model. The resulting estimates turn out to be quite close to the Bank of Russia’s historical forecasts built under comparable conditions.

Multistage optimization filter for trend‐based short‐term forecasting

AbstractA new method is proposed to estimate the long‐term seasonal component by a multistage optimization filter with a leading phase shift (MOPS). It can be utilized to provide better predictions in case of the seasonal component autoregressive (SCAR) model, where data are filtered/decomposed into trend and remainder components and then forecasts for constituent components generated separately and later combined. This reinforces the importance of trend estimation filtering/decomposition methods, which are scarce and only few methods, primarily wavelet decomposition, have improved upon the forecasts generated by statistical linear models. We contribute to the literature by introducing a new trend estimation method, and the forecast results are compared with the most popular trend estimation methods, such as frequency filters, wavelet decomposition, empirical mode decomposition (EMD), and Hodrick–Prescott (HP) filter, through their performance in generating short‐term forecasts for day‐ahead electricity prices. Our method for trend estimation performs better in terms of providing short‐term forecasts as compared with some well‐known methods, and the best forecast, according to the Diebold and Mariano (1995) test, is obtained by using our MOPS filter with annual trend period length.

A modified Diebold–Mariano test for equal forecast accuracy with clustered dependence

This study proposes a modified Diebold–Mariano (DM) test for equal forecast accuracy with clustered dependence. A novel consistent long-run variance estimator is developed to account for the clustered dependence. The modified DM test statistic asymptotically follows a normal distribution. The moving block bootstrap is employed to improve the size and power performance of the newly proposed test. A Monte Carlo simulation shows that the modified DM test has a better finite sample performance than the conventional DM test.

Evaluating probabilistic ecological forecasts.

Probabilistic near-term forecasting facilitates evaluation of model predictions against observations and is of pressing need in ecology to inform environmental decision-making and effect societal change. Despite this imperative, many ecologists are unfamiliar with the widely used tools for evaluating probabilistic forecasts developed in other fields. We address this gap by reviewing the literature on probabilistic forecast evaluation from diverse fields including climatology, economics, and epidemiology. We present established practices for selecting evaluation data (end-sample hold out), graphical forecast evaluation (times-series plots with uncertainty, probability integral transform plots), quantitative evaluation using scoring rules (log, quadratic, spherical, and ranked probability scores), and comparing scores across models (skill score, Diebold-Mariano test). We cover common approaches, highlight mathematical concepts to follow, and note decision points to allow application of general principles to specific forecasting endeavors. We illustrate these approaches with an application to a long-term rodent population time series currently used for ecological forecasting and discuss how ecology can continue to learn from and drive the cross-disciplinary field of forecasting science.

Oil price analysts' forecasts

AbstractCrude oil analysts provide forecasts on future spot prices, which are collected by Bloomberg. We exploit this survey to compare analysts' forecasting ability to futures contracts and also among analysts themselves. We address the problems arising from unstructured forecast data and use the Mean‐Squared Prediction Error (MSPE) relative to the no‐change forecast and the Diebold and Mariano test. The applied approach represents a substantial improvement compared with the standard MSPE methodology as it corrects for volatility and maturity effects on forecasting performance measures. Finally, we establish that futures prices supersede analyst forecasts and elaborate a performance‐based ranking of analyst firms.

Short-Term Electricity Price Forecasting by Employing Ensemble Empirical Mode Decomposition and Extreme Learning Machine

Day-ahead electricity price forecasting plays a critical role in balancing energy consumption and generation, optimizing the decisions of electricity market participants, formulating energy trading strategies, and dispatching independent system operators. Despite the fact that much research on price forecasting has been published in recent years, it remains a difficult task because of the challenging nature of electricity prices that includes seasonality, sharp fluctuations in price, and high volatility. This study presents a three-stage short-term electricity price forecasting model by employing ensemble empirical mode decomposition (EEMD) and extreme learning machine (ELM). In the proposed model, the EEMD is employed to decompose the actual price signals to overcome the non-linear and non-stationary components in the electricity price data. Then, a day-ahead forecasting is performed using the ELM model. We conduct several experiments on real-time data obtained from three different states of the electricity market in Australia, i.e., Queensland, New South Wales, and Victoria. We also implement various deep learning approaches as benchmark methods, i.e., recurrent neural network, multi-layer perception, support vector machine, and ELM. In order to affirm the performance of our proposed and benchmark approaches, this study performs several performance evaluation metric, including the Diebold–Mariano (DM) test. The results from the experiments show the productiveness of our developed model (in terms of higher accuracy) over its counterparts.

Classical and machine learning modeling of crude oil production in Nigeria: Identification of an eminent model for application

The production and exportation of petroleum plays a dominant role in Nigeria’s economy. A drastic fall in the quantity of crude oil produced is fatal to the Nigerian economy since Nigeria depends heavily on the oil sector. Different classical statistical methods have been used to model the crude oil production but the application of machine learning models on crude oil production has been grossly understudied. In this study, we identified a high-performance model between a classical statistical model (ARIMA) and two machine learning models (ANN and RF) in modeling the crude oil production in Nigeria. The monthly data on crude oil production in Nigeria collected from January, 2006 to October, 2020 used in this study is secondary data from Nigerian National Petroleum Corporation (NNPC) and Reuters. The ANN model has a significant modified Diebold–Mariano test when compared with the ARIMA and RF models in the test set. The ANN model had the best trade-off of Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Nash–Sutcliffe Efficiency (NSE) values and was used for forecasting crude oil production in Nigeria. The crude oil production is predicted to increase slightly from March, 2021 (1.67 mmbd) to September, 2021 (1.685 mmbd) before it begins a downward trend from October, 2021 (1.684 mmbd) till it gets to 1.628 mmbd in December, 2023. This predicted fall in the quantity of crude oil produced will obviously affect the Nigerian economy since Nigeria depends solely on the oil sector. There is need to diversify the monoculture economic system in Nigeria to explore and harness untapped natural resources.

Time series forecasting methods for the Baltic dry index

AbstractThis paper forecasts the daily Baltic Dry Index (BDI) using time series and machine learning methods. Significant business cycles and freight rate volatility present in the ocean‐going shipping industry make the ability to forecast freight rates and cycles extremely important for business decisions. Data‐driven model selection based on data characteristics is performed through ARIMA, fractional ARIMA (FARIMA), and ARIMA and FARIMA models with GARCH and EGARCH errors. The corresponding machine learning techniques utilized are feed‐forward fully connected artificial neural networks (ANNs), support vector regression (SVR), and multivariate adaptive regression splines (MARS). Among others, FARIMA and MARS models are used for the first time in forecasting the BDI. Diebold–Mariano tests reveal that time series and machine learning approaches have comparable performance. However, combinations of forecasts of the selected models lead to better forecasting accuracy overall. Bai and Perron tests are utilized to check the robustness of the results over different cycles through the detection of breakpoints in the series.

The forecasting of consumer exchange-traded funds (ETFs) via grey relational analysis (GRA) and artificial neural network (ANN)

Our study uses the grey relational analysis (GRA) and artificial neural network (ANN) models for the prediction of consumer exchange-traded funds (ETFs). We apply eight variables, including the put/call ratio, the EUR/USD exchange rate, the volatility index, the Commodity Research Bureau Index (CRB), the short-term trading index, the New York Stock Exchange Composite Index, inflation, and the interest rate. The GRA model results showed that the NYSE, CRB, EUR/USD, and PCR were the four main variables influencing consumer ETFs. The GRA test results of all the ANN models' data showed that the back propagation neural network (BPN) was the best predictive model. Based on the classification of different percentages of training data, the results of GRA revealed that the radial basis function neural network and the time-delay recurrent neural network exhibited consistent results, compared to BPN and the recurrent neural network. The results also pointed out that different percentages of training data were suitable for predicting consumer ETFs' performance based on high and low grey relationship grade variables. Evidence has shown that the ETFs in Brazil and China are more predictable than those in other countries. All ANN models' results indicated that the use of 10% testing data could predict consumer ETFs better, particularly the ETFs of the United States (US) and those excluding the United States (EX-US). The Diebold–Mariano (DM) test results suggest that the best predictability model for consumer ETFs is BPN, which is significantly superior to other models.

Comparative Analysis of Different Univariate Forecasting Methods in Modelling and Predicting the Romanian Unemployment Rate for the Period 2021-2022.

Unemployment has risen as the economy has shrunk. The coronavirus crisis has affected many sectors in Romania, some companies diminishing or even ceasing their activity. Making forecasts of the unemployment rate has a fundamental impact and importance on future social policy strategies. The aim of the paper is to comparatively analyze the forecast performances of different univariate time series methods with the purpose of providing future predictions of unemployment rate. In order to do that, several forecasting models (seasonal model autoregressive integrated moving average (SARIMA), self-exciting threshold autoregressive (SETAR), Holt–Winters, ETS (error, trend, seasonal), and NNAR (neural network autoregression)) have been applied, and their forecast performances have been evaluated on both the in-sample data covering the period January 2000–December 2017 used for the model identification and estimation and the out-of-sample data covering the last three years, 2018–2020. The forecast of unemployment rate relies on the next two years, 2021–2022. Based on the in-sample forecast assessment of different methods, the forecast measures root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percent error (MAPE) suggested that the multiplicative Holt–Winters model outperforms the other models. For the out-of-sample forecasting performance of models, RMSE and MAE values revealed that the NNAR model has better forecasting performance, while according to MAPE, the SARIMA model registers higher forecast accuracy. The empirical results of the Diebold–Mariano test at one forecast horizon for out-of-sample methods revealed differences in the forecasting performance between SARIMA and NNAR, of which the best model of modeling and forecasting unemployment rate was considered to be the NNAR model.

Forecasting the number of confirmed new cases of COVID-19 in Italy for the period from 19 May to 2 June 2020.

In this paper we forecast the spread of the coronavirus disease 2019 outbreak in Italy in the time window from May 19 to June 2, 2020. In particular, we consider the forecast of the number of new daily confirmed cases. A forecast procedure combining a log-polynomial model together with a first-order integer-valued autoregressive model is proposed. An out-of-sample comparison with forecasts from an autoregressive integrated moving average (ARIMA) model is considered. This comparison indicates that our procedure outperforms the ARIMA model. The Root Mean Square Error (RMSE) of the ARIMA is always greater than that of the our procedure and generally more than twice as high as the our procedure RMSE. We have also conducted Diebold and Mariano (1995) tests of equal mean square error (MSE). The tests results confirm that forecasts from our procedure are significantly more accurate at all horizons. We think that the advantage of our approach comes from the fact that it explicitly takes into account the number of swabs.

Forecasting prices of dairy commodities – a comparison of linear and nonlinear models

Dairy commodity prices have become more volatile over the last 10–11 yr. The aim of this paper was to produce reliable price forecasts for the most frequently traded dairy commodities. Altogether five linear and nonlinear time series models were applied. The analysis reveals that prices of dairy commodities reached a structural breakpoint in 2006/2007. The results also show that a combination of linear and nonlinear models is useful in forecasting commodity prices. In this study, the price of cheese is the most difficult to forecast, but a simple autoregressive (AR) model performs reasonably well after 12 mo. Similarly, for butter the AR model performs the best, while for skimmed milk powder (Smp), whole milk powder (Wmp) and whey powder (Whp) the nonlinear methods are the most accurate. However, few of the differences between models are significant according to the Diebold–Mariano (DM) test. The findings could be of interest to the whole dairy industry.

Bitcoin Analysis and Forecasting through Fuzzy Transform

Sentiment analysis to characterize the properties of Bitcoin prices and their forecasting is here developed thanks to the capability of the Fuzzy Transform (F-transform for short) to capture stylized facts and mutual connections between time series with different natures. The recently proposed Lp-norm F-transform is a powerful and flexible methodology for data analysis, non-parametric smoothing and for fitting and forecasting. Its capabilities are illustrated by empirical analyses concerning Bitcoin prices and Google Trend scores (six years of daily data): we apply the (inverse) F-transform to both time series and, using clustering techniques, we identify stylized facts for Bitcoin prices, based on (local) smoothing and fitting F-transform, and we study their time evolution in terms of a transition matrix. Finally, we examine the dependence of Bitcoin prices on Google Trend scores and we estimate short-term forecasting models; the Diebold–Mariano (DM) test statistics, applied for their significance, shows that sentiment analysis is useful in short-term forecasting of Bitcoin cryptocurrency.