Forecasting trend changes of cement demand in the United States: An exploratory study
This work explores the feasibility of predicting major changes in the trend of cement consumption in the United States based on numerous publicly available economic time series data. The study demonstrates both the challenges and potential inherent in such a forecasting task. A key challenge is the...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-03-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590123024021029 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841555652386750464 |
|---|---|
| author | Ardavan Yazdanbakhsh |
| author_facet | Ardavan Yazdanbakhsh |
| author_sort | Ardavan Yazdanbakhsh |
| collection | DOAJ |
| description | This work explores the feasibility of predicting major changes in the trend of cement consumption in the United States based on numerous publicly available economic time series data. The study demonstrates both the challenges and potential inherent in such a forecasting task. A key challenge is the large number of available predictor time series with different lengths. To address this, a unique segmentation method was developed to handle length discrepancies, and a series synchronization procedure was designed to minimize the lag order of forecast models. Two variable selection techniques were developed for time series of different lengths utilizing lasso and combinatorial optimization. In addition, a novel validation technique was devised using the bootstrap to evaluate models based on the variance of predicted time series values at the most recent known time points. The results revealed significant potential for trend change forecasting as the models correctly predicted the direction of most trend changes. However, the predicted magnitudes of these changes were often smaller than the observed values. Furthermore, the study found that correlations between time series and their relative lags varied over time. Future research should focus on more robust models with higher lag orders that can better adapt to dynamic relationships between time series. |
| format | Article |
| id | doaj-art-0f2274880f79433faa80ab77b91c7e33 |
| institution | Kabale University |
| issn | 2590-1230 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Results in Engineering |
| spelling | doaj-art-0f2274880f79433faa80ab77b91c7e332025-01-08T04:53:24ZengElsevierResults in Engineering2590-12302025-03-0125103859Forecasting trend changes of cement demand in the United States: An exploratory studyArdavan Yazdanbakhsh0Department of Civil Engineering, City College of New York, New York, USAThis work explores the feasibility of predicting major changes in the trend of cement consumption in the United States based on numerous publicly available economic time series data. The study demonstrates both the challenges and potential inherent in such a forecasting task. A key challenge is the large number of available predictor time series with different lengths. To address this, a unique segmentation method was developed to handle length discrepancies, and a series synchronization procedure was designed to minimize the lag order of forecast models. Two variable selection techniques were developed for time series of different lengths utilizing lasso and combinatorial optimization. In addition, a novel validation technique was devised using the bootstrap to evaluate models based on the variance of predicted time series values at the most recent known time points. The results revealed significant potential for trend change forecasting as the models correctly predicted the direction of most trend changes. However, the predicted magnitudes of these changes were often smaller than the observed values. Furthermore, the study found that correlations between time series and their relative lags varied over time. Future research should focus on more robust models with higher lag orders that can better adapt to dynamic relationships between time series.http://www.sciencedirect.com/science/article/pii/S2590123024021029ForecastingTime seriesCementDemandEconomic indicators |
| spellingShingle | Ardavan Yazdanbakhsh Forecasting trend changes of cement demand in the United States: An exploratory study Results in Engineering Forecasting Time series Cement Demand Economic indicators |
| title | Forecasting trend changes of cement demand in the United States: An exploratory study |
| title_full | Forecasting trend changes of cement demand in the United States: An exploratory study |
| title_fullStr | Forecasting trend changes of cement demand in the United States: An exploratory study |
| title_full_unstemmed | Forecasting trend changes of cement demand in the United States: An exploratory study |
| title_short | Forecasting trend changes of cement demand in the United States: An exploratory study |
| title_sort | forecasting trend changes of cement demand in the united states an exploratory study |
| topic | Forecasting Time series Cement Demand Economic indicators |
| url | http://www.sciencedirect.com/science/article/pii/S2590123024021029 |
| work_keys_str_mv | AT ardavanyazdanbakhsh forecastingtrendchangesofcementdemandintheunitedstatesanexploratorystudy |