Development Of Forecasting Model Research Articles

Development of crop yield forecast models under FASAL- a case study of kharif rice in West Bengal

Crop yield forecasts are prepared at District, State and National level under the ongoing project “Forecasting Agricultural output using Space, Agrometeorology and Land based observations (FASAL)”, operational at Ministry of Agriculture, Govt. of India in collaboration with Space Application Centre (SAC), Institute of Economic Growth (IEG) and India Meteorological Department (IMD). As per the mandate of the project, crop yield forecasts are being generated for 13 major crops. Different models are used by these organizations for generation of crop yield forecasts. Under FASAL project, IMD in collaboration with 46 Agromet Field units (AMFU) located at different State Agricultural Universities (SAUs), ICAR institutes, IITs, develops intra-seasonal operational yield forecast for the major crops during kharif and rabi seasons using statistical model. Within IMD, in addition to the Agricultural Meteorology Division, all the Regional Meteorological Centres (RMCs) and Meteorological Centres (MCs), located in different states are also working in this project. Long period crop yield data as well as weekly weather data as per meteorological standard week have been used for development of the district level yield forecast models. For developing the yield forecast, models using composite weather variables have been studied. Simple and weighted weather indices have been prepared for individual weather variables as well as for interaction of two at a time considering throughout the crop growing season. Minimum data set required to develop the statistical model have also been mentioned in this paper. In order to demonstrate the data requirement, methodology for crop yield forecast through regression technique along with the interpretations, the relevant data of West Bengal have been taken up. National level crop yield forecast is prepared by the Mahalanabis National Crop Forecasting Centre (MNCFC) based on the state level forecast generated in IMD.

Erratum for “Purdue Index for Construction Analytics: Prediction and Forecasting Model Development” by Arkaprabha Bhattacharyya, Soojin Yoon, Theodore J. Weidner, and Makarand Hastak

Erratum for “Purdue Index for Construction Analytics: Prediction and Forecasting Model Development” by Arkaprabha Bhattacharyya, Soojin Yoon, Theodore J. Weidner, and Makarand Hastak

Multi-dimensional modeling of radiation belt electron pitch-angle diffusion coefficients caused by plasmaspheric hiss

Plasmaspheric hiss is an important wave mode in the Earth’s radiation belts. Hiss waves can scatter energetic electrons into loss cones to precipitate into the atmosphere, and therefore become an important source of fluctuations, leading the radiation belt to lose electrons . As a function of electron energy and pitch angle, the diffusion coefficient of hiss waves for radiation belt electrons is significantly influenced by the solar wind and geomagnetic activity, and also strongly depends on the spatial position, the background magnetic field, and the plasma density distribution. In order to quickly obtain the diffusion coefficients of hiss waves on electrons in the radiation belt for modelling the global dynamics of the radiation belt, we systematically calculate the diffusion coefficients of hiss waves on electrons in the radiation belt by using the full diffusion code (FDC), and build a four-dimensional matrix database of diffusion coefficients for the spatial region <i>L</i> = 1.5–6, the cold plasma parameter <i>α</i><sup>*</sup> = 3–30, electron energy 1 keV–10 MeV, and electron throw angle 0°–90°. According to the database, we can quickly obtain diffusion coefficients with different <i>L</i> and <i>α</i><sup>*</sup> values through linear interpolations. By comparing the errors between diffusion coefficients calculated by the FDC code and those linearly interpolated from the diffusion coefficient database, the accuracies of interpolated coefficients are validated, showing that most of the errors lie in 10%. The four-dimensional database of hiss wave pitch angle diffusion coefficients for radiation belt electrons and the validated linear interpolation method established in this paper can significantly reduce the time required to obtain global information about hiss wave diffusion coefficients, thereby rapidly improving the computational efficiency of carrying out simulations of spatial and temporal changes in the radiation belts over long periods of time, which in turn is expected to provide favourable conditions for the development of dynamic forecasting models of the Earth's radiation belts.

МЕТОДЫ ПРОГНОЗИРОВАНИЯ МОЩНОСТИ ВЕТРА ДЛЯ ШЕЛЬФОВЫХ ВЕТРОЭЛЕКТРОСТАНЦИЙ

Relevance Wind energy forecasting is an opportunity to evaluate the production possibilities of a wind farm in the short term. Production often refers to the available capacity of the wind farm in question. For example, to date, the installed wind power in Russia has reached 20 GW. Direct transmission operators use existing tools to forecast wind production up to 48 hours. Forecasting tools help optimize power system management. This article discusses the abundance of relevant forecasting methods in the field of wind energy, evaluates their effectiveness and value for the most effective control of wind energy. Particular attention is paid to the ongoing development of wind energy forecasting models to meet the specifics of shelf. Aim of research Conduct a comparative analysis of existing forecasting methods in the field of wind energy under general given conditions, choose the best method for a particular case. Research methods To solve the problem, the authors conducted a comparative analysis of the popular, currently existing methods for forecasting wind farms, comparing their applicability with the specification of the area of use. Results In the course of the study, modern wind energy forecasting tools were analyzed, a comparative analysis was carried out, and conclusions were drawn about the applicability of each of the methods. Keywords: wind energy, short-term forecasting, shelf, optimization, efficiency, model, tool, control, mathematical model, forecast error level

Development of yield forecast models with raw data through stepwise regression analysis (forward method) and discriminant function analysis for pigeon pea crop on Chhattisgarh plain zone

Under this investigation we studied about minimization of weekly data through stratification and Simple Randon Sample technique. Individual effect of weather parameters has been considered under the study. Forecasting model developed on two statistical technique i.e. Stepwise Regression Analysis (Forward Method) technique and Discriminat Function Analysis technique. Both the techniques are found very suitable for the district wise and zone level pre-harvest forecast for the pigeonpea crop on Chhattisgarh. Models fitted with the Stepwise Regression Analysis on 35 variables and Time Trend (T) for Zone. The developed model is showing best fit on the basis of very high value of significance and maximum R2 (84%) for the plain Zone, highly significant at 0.1% level of significance and models fitted with DFA then 2 Discriminant score (ds1 and ds2) and Time Trend (T) the models found highly significant and R2 value 69% for plain zone for Pigeonpea crop. The developed model has been validated by the error parameters viz. Minimum MAE, Minimum MSE, Minimum RMSE, Minimum PE and Minimum PD along with maximum R2. These developed models are useful to farmers of Chhattisgarh to decide their future prospects and possible course of action in advance. This was very challenging task for the researchers to develop a precise & accurate and best fit model for the future forecast. May this study will clear the fog through new methodology and also will create the new way on the direction of forecast.

Development and evaluation of bidirectional LSTM freeway traffic forecasting models using simulation data

Long short-term memory (LSTM) models provide high predictive performance through their ability to recognize longer sequences of time series data. More recently, bidirectional deep learning models (BiLSTM) have extended the LSTM capabilities by training the input data twice in forward and backward directions. In this paper, BiLSTM short term traffic forecasting models have been developed and evaluated using data from a calibrated micro-simulation model for a congested freeway in Melbourne, Australia. The simulation model was extensively calibrated and validated to a high degree of accuracy using field data collected from 55 detectors on the freeway. The base year simulation model was then used to generate loop detector data including speed, flow and occupancy which were used to develop and compare a number of LSTM models for short-term traffic prediction up to 60 min into the future. The modelling results showed that BiLSTM outperformed other predictive models for multiple prediction horizons for base year conditions. The simulation model was then adapted for future year scenarios where the traffic demand was increased by 25–100 percent to reflect potential future increases in traffic demands. The results showed superior performance of BiLSTM for multiple prediction horizons for all traffic variables.

Comparison of Lasso and stepwise regression technique for wheat yield prediction

Multiple regression approach has been used to forecast the crop production widely. This study has been undertaken to evaluate the performance of stepwise and Lasso (Least absolute shrinkage and selection operator) regression technique in variable selection and development of wheat forecast model for crop yield using weather data and wheat yield for the period of 1984-2015, collected from IARI, New Delhi. Statistical parameters viz. R2, RMSE, and MAPE were 0.81, 195.90 and 4.54 per cent respectively with stepwise regression and 0.95, 99.27, 2.7 percentage, respectively with Lasso regression. Forecast models were validated during 2013-14 and 2014-15. Prediction errors were -8.5 and 10.14 per cent with stepwise and 1.89 and 1.64 percent with the Lasso. This shows that performance of Lasso regression is better than stepwise regression to some extent.

Pre-harvest forecast of kharif rice yield using PCA and MLR technique in Navsari district of Gujarat

In this paper Principal Components (PC) and Multiple Linear Regression (MLR) Technique were used for development of pre-harvest model for rice yield in the Navsari district of south Gujarat. The weather indices were developed and utilized for development of pre-harvest forecast models. The data of rice yield and weather parameters from 1990 to 2012 were utilized. The cross validation of the developed forecast model were confirmed using data of the years 2013 to 2016. It was observed that value of Adj. R2 varied from 89 to 96. The appropriate forecast model was selected based on high value of Adj. R2. Based on the outcomes in Navsari district, MLR techniques found to be better than PCA for pre harvest forecasting of rice crop yield. The Model-2 found competent to forecast rice yield in Navsari district before eight weeks of actual harvest of crop (37th SMW) i.e during reproductive stage of the crop growth period.

Development of 30-day mortality forecast model in patients after surgical treatment of proximal hip fracture

Introduction Fractures of the proximal femur often occur in elderly and senile patients. Most of them have an unfavorable comorbid background. In this regard, there is a high risk of complications in the postoperative period, which requires the development and implementation of an effective forecasting model in clinical practice designed to take measures to avoid adverse treatment outcomes.Objective To develop a regression model to predict the probability of lethal outcome within 30 days after surgery in patients with proximal femur fracture.Material and Methods A retrospective analysis of inpatient case histories of all patients (n = 1222) with proximal femur fracture treated in our hospital in 2018-2019 was performed. A total of 388 cases were selected for the study.Results After a detailed statistical analysis of the physiological parameters of the patients, four independent factors were identified that increased the risk of death during 30 days following surgery: albumin less than 30 g / l (regression coefficient – 1.742; OR – 5.708, 95% CI – 1.904–17.114, p = 0.002), the presence of diabetes mellitus (regression coefficient – 1.141; OR – 3.130, 95% CI –1.022–9.588, p = 0.046), the presence of acute renal injury (regression coefficient – 3.141; OR – 23.136, 95% CI – 3.886–137.735, p = 0.001), the presence of pneumonia (regression coefficient – 2.130; OR – 8.411, 95% CI – 2.453–28.838, p = 0.001). A regression model for predicting 30-day mortality was developed: the constant regression coefficient was 4.371, the area under the ROC-curve corresponding to the probability of 30-day mortality was 0.841 with 95% CI: 0.732–0.951, model sensitivity and specificity – 78.9 and 81.2%, respectively. After a detailed statistical analysis of the patients' physiological parameters, four independent factors were identified that increase the risk of fatal outcome during the next 30 days after surgery: albumin less than 30 g / l (regression coefficient, 1.742; OR – 5.708, 95% CI (1.904 – 17.114), p = 0.002), presence of diabetes mellitus (regression coefficient – 1.141; OR – 3.130, 95% CI (1.022 – 9.588), p = 0.046), presence of acute renal injury (regression coefficient – 3.141; OR – 23.136, 95% CI (3.886 – 137.735), p = 0.001), presence of pneumonia (regression coefficient – 2.130; OR – 8.411, 95% CI (2.453 – 28.838), p = 0.001). Thereby we developed a regression model to predict 30-day mortality: regression coefficient of the constant was 4.371; area under the ROC curve, corresponding to the dependence of the probability of 30-day mortality, was 0.841 with 95% CI (0.732 – 0.951); model sensitivity and specificity were 78.9 and 81.2%, respectively.Conclusion Aregression model for predicting mortality in patients with proximal femur fractures based on independent risk factors has a sufficient level of sensitivity and specificity. Its application is possible in practical health care institutions, where patients with trauma are treated.

Purdue Index for Construction Analytics: Prediction and Forecasting Model Development

The Purdue Index for Construction (Pi-C) was developed to gauge the health of the construction industry. It is a composite index consisting of five dimensions: economic, stability, social, development, and quality. This research conducts a data-driven analysis to provide prediction and time-series forecasting models for Pi-C to (1) monitor; and (2) provide guidance on how to improve the future health trajectory for the US construction industry. The seasonal autoregressive integrated moving average (SARIMA) technique is applied for the future trend analysis; multiple linear regression (MLR) and random forests (RF) are applied for prediction models of Pi-C data analytics. It is expected that the proposed prediction and time-series forecasting models will help decision-makers, including policy developers and construction practitioners, to take necessary action in a timely manner, as well as open the discourse on the advanced application of analytics and data-driven decision-making in the construction industry.

Development of real-time flood forecast model for vamsadhara river through hydrological approach

Development of real-time flood forecast model for vamsadhara river through hydrological approach

Manpower forecasting models in the construction industry: a systematic review

PurposeThis paper aims to make a systematic review of the manpower prediction model of the construction industry. It aims to determine the forecasting model's development trend, analyse the use limitations and applicable conditions of each forecasting model and then identify the impact indicators of the human resource forecasting model from an economic point of view. It is hoped that this study will provide insights into the selection of forecasting models for governments and groups that are dealing with human resource forecasts.Design/methodology/approachThe common search engine, Scopus, was used to retrieve construction manpower forecast-related articles for this review. Keywords such as “construction”, “building”, “labour”, “manpower” were searched. Papers that not related to the manpower prediction model of the construction industry were excluded. A total of 27 articles were obtained and rated according to the publication time, author and organisation of the article. The prediction model used in the selected paper was analysed.FindingsThe number of papers focussing on the prediction of manpower in the construction industry is on the rise. Hong Kong is the region with the largest number of published papers. Different methods have different requirements for the quality of historical data. Most forecasting methods are not suitable for sudden changes in the labour market. This paper also finds that the construction output is the economic indicator with the most significant influence on the forecasting model.Research limitations/implicationsThe research results discuss the problem that the prediction results are not accurate due to the sudden change of data in the current prediction model. Besides, the study results take stock of the published literature and can provide an overall understanding of the forecasting methods of human resources in the construction industry.Practical implicationsThrough this study, decision-makers can choose a reasonable prediction model according to their situation. Decision-makers can make clear plans for future construction projects specifically when there are changes in the labour market caused by emergencies. Also, this study can help decision-makers understand the current research trend of human resources forecasting models.Originality/valueAlthough the human resource prediction model's effectiveness in the construction industry is affected by the dynamic change of data, the research results show that it is expected to solve the problem using artificial intelligence. No one has researched this area, and it is expected to become the focus of research in the future.

Performance evaluation of Emergency Department patient arrivals forecasting models by including meteorological and calendar information: A comparative study

The volume of daily patient arrivals at Emergency Departments (EDs) is unpredictable and is a significant reason of ED crowding in hospitals worldwide. Timely forecast of patients arriving at ED can help the hospital management in early planning and avoiding of overcrowding. Many different ED patient arrivals forecasting models have previously been proposed by using time series analysis methods. Even though the time series methods such as Linear and Logistic Regression, Autoregressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), Exponential Smoothing (ES), and Artificial Neural Network (ANN) have been explored extensively for the ED forecasting model development, the few significant limitations of these methods associated in the analysis of time series data make the models inadequate in many practical situations. Therefore, in this paper, Machine Learning (ML)-based Random Forest (RF) regressor, and Deep Neural Network (DNN)-based Long Short-Term Memory (LSTM) and Convolutional Neural network (CNN) methods, which have not been explored to the same extent as the other time series techniques, are implemented by incorporating meteorological and calendar parameters for the development of forecasting models. The performances of the developed three models in forecasting ED patient arrivals are evaluated. Among the three models, CNN outperformed for short-term (3 days in advance) patient arrivals prediction with Mean Absolute Percentage Error (MAPE) of 9.24% and LSTM performed better for moderate-term (7 days in advance) patient arrivals prediction with MAPE of 8.91% using weather forecast information. Whereas, LSTM model outperformed with MAPE of 8.04% compared to 9.53% by CNN and 10.10% by RF model for current day prediction of patient arrivals using 3 days past weather information. Thus, for short-term ED patient arrivals forecasting, DNN-based model performed better compared to RF regressor ML-based model.

Development of mathematical models for the short-term forecasting of daily consumption schedules of active power by Moscow

The present article presents the results of research into solving the problem of increasing the accuracy of forecasting power consumption. The purpose of these studies is to develop mathematical models for short-term forecasting of daily schedules of active power consumption in Moscow, taking into account meteorological factors. Research has been carried out on four predictive models based on singular spectral analysis (SSA), least-squares method, trigonometric interpolation, neural and neural fuzzy networks (NFN). It is shown that the NFN and hybrid model based on MSSA and NFN has the smallest error.

Managing Pandemics with Health Informatics: Successes and Challenges.

SummaryIntroduction : The novel COVID-19 pandemic struck the world unprepared. This keynote outlines challenges and successes using data to inform providers, government officials, hospitals, and patients in a pandemic. Methods : The authors outline the data required to manage a novel pandemic including their potential uses by governments, public health organizations, and individuals. Results : An extensive discussion on data quality and on obstacles to collecting data is followed by examples of successes in clinical care, contact tracing, and forecasting. Generic local forecast model development is reviewed followed by ethical consideration around pandemic data. We leave the reader with thoughts on the next inevitable outbreak and lessons learned from the COVID-19 pandemic. Conclusion : COVID-19 must be a lesson for the future to direct us to better planning and preparing to manage the next pandemic with health informatics.

Development of fine-scale spatiotemporal temperature forecast model with urban climatology and geomorphometry in Hong Kong

High precisive fine-scale temperature forecast serves as an important tool in quantifying human comfort, health threats with extreme temperature and energy consumption. In this study, Urban Microscale Temperature Forecast (UMTF) model was developed using machine learning techniques (k-means clustering and support vectors machine) with reference to global ensembles and geomorphometry datasets (sky view factor, daily sun trajectory and urban terrain model). It statistically downscales the daily temperature extremes of 10-km ensemble forecast into 50-m resolution across Hong Kong. It is observed that the downscaling process has contributed to a considerable reduction of mean absolute errors (up to 1.0 degree Celsius (45.1%)) with improved correlation (R2 = 0.84) across 26 automatic weather stations. To further validate the result, an independent dataset from the Community Weather Information Network (CoWIN) and the ASTER heat island intensity map were added for the model evaluation. The comparison result illustrates clear and resemble spatiality in temperature extremes with sophisticated urban buildings, park greenery and coastal front features can be re-projected in details. At last, the study demonstrates the low-cost weather stations could be an effective way to achieve a detailed temperature forecast in a heterogeneous urban environment.

Modelling hybrid and backpropagation adaptive neuro-fuzzy inference systems for flood forecasting

The ability of the adaptive neuro-fuzzy inference algorithm architecture to simulate floods is explored in this research. The development of models for flood forecasting has been centered on two adaptive neuro-fuzzy inference (ANFIS) algorithms. The Takagi–Sugeno fuzzy inference systems (FIS) generated through subtracted clustering were trained using hybrid and backpropagation training algorithms. Multiple statistical performance evaluators were used to assess the performability of the established models. The validity and predictive power of the models are evaluated by estimating a flood occurrence in the study area. In designing the models, a total of 12 inputs were employed. The best performability was found for the ANFIS model created utilizing a hybrid training algorithm with mean square error (MSE) of 0.00034, co-efficient of correlation (R2) of 97.066%, root mean square error (RMSE) of 0.018, Nash–Sutcliffe model efficiency (NSE) of 0.968, mean absolute error (MAE) of 0.0073 and combined accuracy (CA) of 0.018, indicating the possible usage of exploiting the established model for prediction of floods.

Sensitivity of QBO teleconnection to model circulation biases

AbstractThe Quasi‐biennial oscillation (QBO) in the equatorial stratospheric winds is known to modulate the extratropical atmospheric circulation so that an easterly QBO phase is associated with a weaker Arctic stratospheric vortex and a negative North Atlantic Oscillation in winter. The link can potentially be used to improve seasonal forecasts and is usually captured by climate models but with a smaller magnitude than that in observations. Here, the sensitivity of the QBO teleconnection to biases in model circulation is explored in an atmosphere‐only model ECHAM6 with an internally generated QBO, by contrasting simulations by an original, biased model, and a model version in which biases are artificially reduced. We find that the strength of the winter mean Arctic stratospheric vortex response is better reproduced by the bias‐corrected model. On the other hand, the response is delayed by 1 month and there is no improvement in the tropospheric response. Stronger response of the polar vortex in the bias‐corrected model is likely associated with larger magnitude of QBO winds and better‐represented stratospheric mean climate and planetary wave fluxes, but the exact mechanism remains unclear. The results of this study have implications for seasonal forecast model development.

Modelling the large earthquakes recurrence times along the North Aegean Trough Fault Zone (Greece) with a physics-based simulator

SUMMARYThe recurrence time of large earthquakes above a predefined magnitude threshold on specific faults or fault segments is one of the key parameters for the development of long-term Earthquake Rupture Forecast models. Observational data of successive large earthquakes per fault segment are often limited and thus inadequate for the construction of robust statistical models. The physics-based earthquake simulators are a powerful tool to overcome recurrence data limitations by generating long earthquake records. A physics-based simulator, embodying well known physical processes, is applied in the North Aegean Trough (NAT) Fault Zone (Greece). The application of the simulation is implemented, after defining a five segment source model, aiming at the investigation of the recurrence behaviour of earthquakes with Mw ≥ 6.5 and Mw ≥ 7.0. The detailed examination of the 544 Mw ≥ 6.5 earthquakes included in the simulated catalogue reveals that both single and multiple segmented ruptures can be realized along the NAT. Results of statistical analysis of the interevent times of Mw ≥ 6.5 and Mw≥ 7.0 earthquakes per participating segment to the related ruptures indicate the better performance of the Brownian Passage Time renewal model in comparison to exponential model. These results provide evidence for quasi-periodic recurrence behaviour, agreeing with the elastic rebound theory, instead of Poissonian behaviour.

Prediction of heat waves over Pakistan using support vector machine algorithm in the context of climate change

Many efficient forecasting models have been found to fail or show low skill due to the changes in the predictor–predictand relationship with the changes in global climate. An attempt has been taken to develop a climate change resilient heatwave prediction model using machine learning (ML) algorithms known as Support Vector Machines (SVM), random forest and artificial neural network. The National Centres for Environmental Prediction/National Centre for Atmospheric Research reanalysis data of ocean-atmospheric variables were used as the predictors of ML models for forecasting the number of heatwave days (HWDs) in the summer of Pakistan. An SVM based recursive feature elimination method was used to select the skilful predictors. The ML models were developed by considering a moving window of 29 years with a time step of 5 years to incorporate the changes in the relation of HWDs with its predictors due to climate change. The result showed changes in the relationship of HWDs with all the ocean-atmospheric variables considered in this study as probable predictors, which indicates the necessity of forward-rolling approach proposed in this study for the development of climate change resilient forecasting model. The relative performance of ML showed the higher capability of SVM to predict HWDs with an %NRMSE of 36, R2 of 0.87, md score of 0.76 and an rSD of 0.88 during the validation period. The result revealed the potential of SVM model to be used for reliable forecasting of heatwaves in the context of climate change.