In this article, we propose the time-series average of spatial HAC estimators for the variance of the estimated common factors under the approximate factor structure. Based on this, we provide the confidence interval for the conditional mean of the diffusion-index forecasting model with cross-sectional heteroscedasticity and dependence of the idiosyncratic errors. We establish the asymptotics under very mild conditions, and no prior information about the dependence structure is needed to implement our procedure. We employ a bootstrap to select the bandwidth parameter. Simulation studies show that our procedure performs well in finite samples. We apply the proposed confidence interval to the problem of forecasting the unemployment rate using data by Ludvigson and Ng.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

VN 30 Index Forecast: Mixed Signals Ahead

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

Deterministic and stochastic are two methods for modeling of crude oil and bottled water market. Forecasting the price of the market directly affected energy producer and water user.There are two software, Tableau and Python, which are utilized to model and visualize both markets for the aim of estimating possible price in the future.The role of those software is to provide an optimal alternative with different methods (deterministic versus stochastic). The base of predicted price in Tableau is deterministic—global optimization and time series. In contrast, Monte Carlo simulation as a stochastic method is modeled by Python software. The purpose of the project is, first, to predict the price of crude oil and bottled water with stochastic (Monte Carlo simulation) and deterministic (Tableau software),second, to compare the prices in a case study of Crude Oil Prices: West Texas Intermediate (WTI) and the U.S. bottled water. 1. Introduction Predicting stock and stock price index is challenging due to uncertainties involved. We can analyze with a different aspect; the investors perform before investing in a stock or the evaluation of stocks by means of studying statistics generated by market activity such as past prices and volumes. The data analysis attempt to identify stock patterns and trends that may predict the estimation price in the future. Initially, the classical regression (deterministic) methods were used to predict stock trends; furthermore, the uncertainty (stochastic) methods were used to forecast as same as deterministic. According to Deterministic versus stochastic volatility: implications for option pricing models (1997), Paul Brockman & Mustafa Chowdhury researched that the stock return volatility is deterministic or stochastic. They reported that “Results reported herein add support to the growing literature on preference-based stochastic volatility models and generally reject the notion of deterministic volatility” (Pag.499). For this argument, we need to research for modeling forecasting historical data with two software (Tableau and Python). In order to forecast analyze Tableau feature, the software automatically chooses the best of up to eight models which generates the highest quality forecast. According to the manual of Tableau , Tableau assesses forecast quality optimize the smoothing of each model. The optimization model is global. The main part of the model is a taxonomy of exponential smoothing that analyzes the best eight models with enough data. The real- world data generating process is a part of the forecast feature and to support deterministic method. Therefore, Tableau forecast feature is illustrated the best possible price in the future by deterministic (time – series and prices). Monte Carlo simulation (MCs) is modeled by Python, which is predicted the floating stock market index . Forecasting the stock market by Monte Carlo demonstrates in mathematics to solve various problems by generating suitable random numbers and observing that fraction of the numbers that obeys some property or properties. The method utilizes to obtain numerical solutions to problems too complicated to solve analytically. It randomly generates thousands of series representing potential outcomes for possible returns. Therefore, the variable price is the base of a random number between possible spot price between 2002-2016 that present a stochastic method.

Outage prediction tool that correlates historical weather data, weather forecast output from the GFS 003 WRF model, and historical outage data to create a 96-hour forecast predicting the potential for weather-related outages and the corresponding number of customers potentially impacted in each of SDG&E’s eight operating districts.

MATERIALS AND INFORMATION ARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT WILL SAN DIEGO GAS & ELECTRIC COMPANY BE LIABLE TO ANY PARTY FOR ANY DIRECT, INDIRECT, SPECIAL OR CONSEQUENTIAL DAMAGES FOR ANY USE OF THE MATERIALS OR INFORMATION PROVIDED HEREIN, INCLUDING, WITHOUT LIMITATION, ANY CLAIMS OR DEMANDS FOR LOST PROFITS OR BUSINESS INTERRUPTION, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

Should You Buy, Sell, or Hold? (SET Index Stock Forecast)

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

BackgroundIn China, dengue remains an important public health issue with expanded areas and increased incidence recently. Accurate and timely forecasts of dengue incidence in China are still lacking. We aimed to use the state-of-the-art machine learning algorithms to develop an accurate predictive model of dengue.Methodology/Principal findingsWeekly dengue cases, Baidu search queries and climate factors (mean temperature, relative humidity and rainfall) during 2011–2014 in Guangdong were gathered. A dengue search index was constructed for developing the predictive models in combination with climate factors. The observed year and week were also included in the models to control for the long-term trend and seasonality. Several machine learning algorithms, including the support vector regression (SVR) algorithm, step-down linear regression model, gradient boosted regression tree algorithm (GBM), negative binomial regression model (NBM), least absolute shrinkage and selection operator (LASSO) linear regression model and generalized additive model (GAM), were used as candidate models to predict dengue incidence. Performance and goodness of fit of the models were assessed using the root-mean-square error (RMSE) and R-squared measures. The residuals of the models were examined using the autocorrelation and partial autocorrelation function analyses to check the validity of the models. The models were further validated using dengue surveillance data from five other provinces. The epidemics during the last 12 weeks and the peak of the 2014 large outbreak were accurately forecasted by the SVR model selected by a cross-validation technique. Moreover, the SVR model had the consistently smallest prediction error rates for tracking the dynamics of dengue and forecasting the outbreaks in other areas in China.Conclusion and significanceThe proposed SVR model achieved a superior performance in comparison with other forecasting techniques assessed in this study. The findings can help the government and community respond early to dengue epidemics.

📦 Dataset Description This dataset provides monthly sales data for industrial pumps and valves, specifically focusing on External Gear Pumps and Valves across various product variants, customer segments, and regions.

It is designed to support demand forecasting, marketing impact analysis, and supply chain optimization use cases.

The dataset combines internal sales metrics (e.g., units sold, revenue, discount, stock) with external factors such as:

Marketing Spend

Competitor Activity

Economic Indicators

New Product Launch Flags

Seasonality Index

These features make it ideal for developing machine learning models that go beyond traditional time series forecasting by incorporating real-world business drivers.

🔍 Key Features Time Period: Monthly data (example: 2020-01 onward)

Granularity: By product variant, region, customer segment, and sales channel

Target Variable: units_sold

Support Variables:

Marketing: marketing_spend, discount_percent

Customer: customer_segment, channel, return_units

Product: variant, pressure_rating_bar, oil_type

External: competitor_activity, economic_indicator, seasonality_ind

This dataset includes model output from the MeteoFrance System8 (MF8) seasonal forecasts. The time period spans the years 1994-2018 for the Alaska fire season (April 1 - September 30) from March-initialized and May-initialized seasonal forecasts. MF8 model data were collected from the Copernicus Climate Data Store (https://cds.climate.copernicus.eu/cdsapp#!/dataset/seasonal-monthly-single-levels?tab=form) for the years 1994-2018. Variables downloaded include 2 meter temperature, 2 meter dew point temperature, precipitation, 10 meter wind speeds, and sea level pressure. These models consist of 25 ensembles. Details about the model can be found in its reference publication: Batté, L., Dorel, L., Ardilouze, C., & Guérémy, J.-F. (2021). Documentation of the METEO-FRANCE seasonal forecasting system 8. http://www.umr-cnrm.fr/IMG/pdf/system8-technical.pdf Model data were extracted for 13 Predictive Service Areas (PSA) in Alaska at 0000 UTC (the time closest to 1400 AKDT) and saved as a weighted area-average in text files by year. Buildup index was calculated from the area-averaged variables and saved in annual text files. Calculations were done with NCAR Command Line (NCL) version 6.3.0. Sub-directory Naming Convention The MF8 directory is divided into 15 sub-directories: one sub-directory for each of the 13 PSAs in this study (named for each PSA), one sub-directory for March-initialized forecast data (step1_mf8data_march), and one sub-directory for May-initialized forecast data (step1_mf8data_may). The step1_mf8data_march and step1_mf8data_may folders contain raw model data from MeteoFrance System8 March-initialized and May-initialized seasonal forecasts, respectively, saved by ensemble in one file for the calculation of buildup index. Data in each file include precipitation (precip), surface pressure (P), 2 meter dewpoint temperature (DP), 2 meter air temperature (T), and 10 meter wind speeds (UV). Data were subset for the state of Alaska and for the Alaska fire season (months of April, May, June, July, August, and September) and saved by year. There are 26 files at a size of 158MB (25 ensembles) for March-initialized forecasts and 26 files at a size of 132MB (25 ensembles) for May-initialized forecasts. Each PSA sub-directory (PSA AKXX) contains 4 additional sub-directories: step2_mf8data_march step2_mf8data_may step3_mf8data_march_BUI step3_mf8data_may_BUI where 'march' or 'may' denotes data from March-initialized forecasts of May-initialized forecasts, respectively. The step2 folders contain precipitation (precip), surface pressure (press), 2 meter dew point temperature (dp), relative humidity (rh), 2 meter air temperature (tmp2m), and 10 meter wind speeds (uv).. These data were extracted from the data files in the step1 sub-directories and saved as annual files. The step3 folders contain the resulting BUI values saved as annual files.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

The Dow Jones U.S. Completion Total Stock Market Index

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

Context

Stocks and financial instrument trading is a lucrative proposition. Stock markets across the world facilitate such trades and thus wealth exchanges hands. Stock prices move up and down all the time and having ability to predict its movement has immense potential to make one rich. Stock price prediction has kept people interested from a long time. There are hypothesis like the Efficient Market Hypothesis, which says that it is almost impossible to beat the market consistently and there are others which disagree with it.

There are a number of known approaches and new research going on to find the magic formula to make you rich. One of the traditional methods is the time series forecasting. Fundamental analysis is another method where numerous performance ratios are analyzed to assess a given stock. On the emerging front, there are neural networks, genetic algorithms, and ensembling techniques.

Another challenging problem in stock price prediction is Black Swan Event, unpredictable events that cause stock market turbulence. These are events that occur from time to time, are unpredictable and often come with little or no warning.

A black swan event is an event that is completely unexpected and cannot be predicted. Unexpected events are generally referred to as black swans when they have significant consequences, though an event with few consequences might also be a black swan event. It may or may not be possible to provide explanations for the occurrence after the fact – but not before. In complex systems, like economies, markets and weather systems, there are often several causes. After such an event, many of the explanations for its occurrence will be overly simplistic.

#
#

https://www.visualcapitalist.com/wp-content/uploads/2020/03/mm3_black_swan_events_shareable.jpg"> #
#
New bleeding age state-of-the-art deep learning models stock predictions is overcoming such obstacles e.g. "Transformer and Time Embeddings". An objectives are to apply these novel models to forecast stock price.

Content

Stock price prediction is the task of forecasting the future value of a given stock. Given the historical daily close price for S&P 500 Index, prepare and compare forecasting solutions. S&P 500 or Standard and Poor's 500 index is an index comprising of 500 stocks from different sectors of US economy and is an indicator of US equities. Other such indices are the Dow 30, NIFTY 50, Nikkei 225, etc. For the purpose of understanding, we are utilizing S&P500 index, concepts, and knowledge can be applied to other stocks as well.

Dataset

The historical stock price information is also publicly available. For our current use case, we will utilize the pandas_datareader library to get the required S&P 500 index history using Yahoo Finance databases. We utilize the closing price information from the dataset available though other information such as opening price, adjusted closing price, etc., are also available. We prepare a utility function get_raw_data() to extract required information in a pandas dataframe. The function takes index ticker name as input. For S&P 500 index, the ticker name is ^GSPC. The following snippet uses the utility function to get the required data.(See Simple LSTM Regression)

Features and Terminology: In stock trading, the high and low refer to the maximum and minimum prices in a given time period. Open and close are the prices at which a stock began and ended trading in the same period. Volume is the total amount of trading activity. Adjusted values factor in corporate actions such as dividends, stock splits, and new share issuance.

Starter Kernel(s)

• Simple LSTM Regression

Acknowledgements

Mining and updating of this dateset will depend upon Yahoo Finance .

Inspiration

Sort of variation of sequence modeling and bleeding age e.g. attention can be applied for research and forecasting

Some Readings

• Applications of deep learning in stock market prediction: recent progress
• Stock predictions with state-of-the-art Transformer and Time Embeddings
• https://arxiv.org/pdf/1712.02136.pdf
• https://catanacapital.com/blog/unpredictable-black-swan-event-stock-market/
• https://par.nsf.gov/servlets/purl/10149715
• https://arxiv.org/pdf/1712.02136.pdf

*If you download and find the data useful your upvote is an explicit feedback for future works*

Number of nowcasting and forecasting models selected in the MCS at the 90% confidence level, using the statistic and the MSE loss function, as well as number of selected models using the “false” Google Index.

Snapshot img

Weather Forecasting Services Market Size 2025-2029

The weather forecasting services market size is valued to increase USD 1.6 billion, at a CAGR of 11.8% from 2024 to 2029. Farmers need weather forecasting services will drive the weather forecasting services market.

Major Market Trends & Insights

APAC dominated the market and accounted for a 40% growth during the forecast period.
By Type - Medium-range segment was valued at USD 555.80 billion in 2023
By Application - Energy and utilities segment accounted for the largest market revenue share in 2023

Market Size & Forecast

Market Opportunities: USD 148.85 million
Market Future Opportunities: USD 1595.10 million
CAGR : 11.8%
APAC: Largest market in 2023

Market Summary

The market encompasses a dynamic and essential industry, driven by advancements in core technologies and applications. With the increasing reliance on accurate weather information for various sectors, such as agriculture and renewable energy production, the market's significance continues to grow. For instance, farmers heavily depend on weather forecasting services to optimize crop yields and mitigate potential losses. Moreover, the upsurge in the production of renewable energy necessitates precise weather predictions to ensure efficient energy generation. However, the complexities of weather forecasting models pose significant challenges. These models must account for numerous variables and continually adapt to evolving weather patterns.
One of the major drivers for the market's growth is the increasing adoption of advanced technologies like Artificial Intelligence (AI) and Machine Learning (ML) to improve forecasting accuracy. As of 2021, AI and ML technologies are estimated to account for over 20% of the market share. Despite these advancements, regulatory frameworks and data privacy concerns pose challenges for market growth. Additionally, regional differences in weather patterns and climatic conditions create diverse market opportunities. As the market continues to evolve, stakeholders must navigate these challenges and capitalize on emerging opportunities to remain competitive.

What will be the Size of the Weather Forecasting Services Market during the forecast period?

Get Key Insights on Market Forecast (PDF) Request Free Sample

How is the Weather Forecasting Services Market Segmented and what are the key trends of market segmentation?

The weather forecasting services industry research report provides comprehensive data (region-wise segment analysis), with forecasts and estimates in 'USD million' for the period 2025-2029, as well as historical data from 2019-2023 for the following segments.

Type

  Medium-range
  Long-range
  Short-range
  Nowcasting


Application

  Energy and utilities
  Aviation
  Media and consumer
  Logistics and transportation
  Others


Method

  Ground-based
  Satellite-based
  Model-based


Geography

  North America

    US
    Canada


  Europe

    France
    Germany
    Italy
    UK


  APAC

    China
    India
    Japan
    South Korea


  Rest of World (ROW)

By Type Insights

The medium-range segment is estimated to witness significant growth during the forecast period.

Weather forecasting services play a crucial role in various sectors, including aviation, agriculture, energy, and finance. Ensemble prediction systems analyze multiple forecasts to enhance accuracy, while aviation weather briefings ensure safe flights. Forecasting model validation ensures reliability, and climate prediction systems help understand long-term trends. Weather risk management mitigates potential losses, and air quality forecasting protects public health. Atmospheric data assimilation combines observations and models, and atmospheric circulation patterns provide context. Weather station networks collect essential data, and severe weather warnings save lives. Weather model ensembles offer probabilistic forecasts, and satellite meteorology provides global coverage. UV index prediction safeguards outdoor activities, and weather prediction accuracy depends on data quality control.

Wind energy forecasting optimizes production, and climate change impacts require adaptation strategies. Marine weather forecasts ensure safe maritime travel, and hydrological forecasting manages water resources. Climate modeling techniques explore future scenarios, high-resolution forecasting enhances precision, and agricultural weather services optimize crop yields. Radar meteorology monitors precipitation, numerical weather prediction models simulate weather, and short-range forecasts provide immediate insights. Geospatial weather data offers location-specific information, and extreme weather events require robust response plans. Model output statistics inform decision-making, and long-range forecasting anticipates trends. Mesoscale modeling focuses on local weather pat

The large-scale Land-Uses and Land-Cover Changes (LULCC) in India in the past several decades is primarily driven by anthropogenic factors that influence the climate from regional to global scales. Therefore, to understand the LULCC over the Indian region from 2002 to 2015 and its implications on temperature and precipitation, we performed Weather Research Forecast (WRF) model simulation using the European Centre for Medium-Range Weather Forecast (ECMWF) reanalysis data for the period 2009 to 2015 as a boundary condition with 2009 as spin-up time. The results showed moderate forest cover loss in major parts of northeast India, and the Himalayan region during 2002–2015. Such large LULC changes, primarily significant alteration of grassland and agriculture from the forest, led to increased precipitation due to increasing evapotranspiration (ET) similar to the forest-dominated regions. An increase in the precipitation patterns (>300 mm) was observed in the parts of eastern and western Himalayas, western Ghats, and the northwestern part of central India, while most parts of northeast Himalayas have an exceptional increase in precipitation (∼100–150 mm), which shows similar agreement with an increase of leaf area index (LAI) by ∼15%. The overall phenomenon leads to a greening-induced ET enhancement that increases atmospheric water vapor content and promotes downwind precipitation. In the case of temperature, warming was observed in the central to eastern parts of India, while cooling was observed in the central and western parts. The increase in vegetated areas over northwest India led to an increase in ET, which ultimately resulted in decreased temperature and increased precipitation. The study highlights the changes in temperature and precipitation in recent decades because of large LULCC and necessitates the formulation of sustainable land use-based strategies to control meteorological variability and augment ecological sustainability.

This dataset includes model output from the NOAA Climate Forecast System version 2 (CFSv2) seasonal forecasts. The time period spans the years 1994-2019 for the Alaska fire season (April 1 - September 30) from March-initialized and May-initialized seasonal forecasts. NOAA CFSv2 model data were collected from the NOAA CFS reforecasts (years 1994- March 2011) and operational forecasts (years April 2011-2019) via the NCEI direct download webpages: https://www.ncei.noaa.gov/oa/prod-cfs-reforecast/first-look/6-hourly-time-series-9-month and https://www.ncei.noaa.gov/data/climate-forecast-system/access/operational-9-month-forecast/. Variables downloaded include 2 meter temperature, 2 meter specific humidity, precipitation rate, 10 meter wind speeds, and surface pressure. This model consists of 25-125 ensembles depending on the availability of data. Details about the model can be found in its reference publication: Saha, S., Moorthi, S., Wu, X., Wang, J., Nadiga, S., Tripp, P., Behringer, D., Hou, Y. T., Chuang, H. Y., Iredell, M., Ek, M., Meng, J., Yang, R., Mendez, M. P., Van Den Dool, H., Zhang, Q., Wang, W., Chen, M., & Becker, E. (2014). The NCEP Climate Forecast System Version 2. Journal of Climate, 27(6), 2185–2208. https://doi.org/10.1175/JCLI-D-12-00823.1 Model data were extracted for 13 Predictive Service Areas (PSA) in Alaska at 0000 UTC (the time closest to 1400 AKDT) and saved as a weighted area-average in text files by year. Buildup index was calculated from the area-averaged variables and saved in annual text files. Calculations were done with NCAR Command Language (NCL) version 6.3.0. Sub-directory Naming Convention The CFSv2 directory is divided into 15 sub-directories: one sub-directory for each of the 13 PSAs in this study (named for each PSA), one sub-directory for March-initialized forecast data (step1_cfsdata_march), and one sub-directory for May-initialized forecast data (step1_cfsdata_may). The step1_cfsdata_march and step1_cfsdata_may folders contain raw model data from NOAA CFSv2 March-initialized and May-initialized seasonal forecasts, respectively, saved by ensemble in one file for the calculation of buildup index. Data in each file include precipitation (precip), surface pressure (P), 2 meter specific humidity (Q), 2 meter air temperature (T), and 10 meter wind speeds (UV). Data were subset for the state of Alaska and for the Alaska fire season (months of April, May, June, July, August, and September) and saved as one file for each model ensemble. There are 1422 files at a size of 7MB for March-initialized forecasts and 1565 files at a size of 5-6MB for May-initialized forecasts. Each PSA sub-directory (PSA AKXX) contains 4 additional sub-directories: step2_cfsdata_march step2_cfsdata_may step3_cfsdata_march_BUI step3_cfsdata_may_BUI where 'march' or 'may' denotes data from March-initialized forecasts of May-initialized forecasts, respectively. The step2 folders contain precipitation (precip), surface pressure (press), 2 meter specific humidity (q), relative humidity (rh), 2 meter air temperature (tmp2m), and 10 meter wind speeds (uv). These data were extracted from the data files in the step1 sub-directories and saved as annual files. The step3 folders contain the resulting BUI values saved as annual files.

The heteroscedastic and volatile characteristics of stock price data have attracted the interest of researchers from various disciplines, particularly in the realm of price forecasting. The stock market’s non-stationary and volatile nature, driven by complex interrelationships among financial assets, economic developments, and market participants, poses significant challenges for accurate forecasting. This research aims to develop a robust forecasting model to improve the accuracy and reliability of stock price predictions using machine learning. A two-stage forecasting model is introduced. First, a random forest subset-based (RFS) feature selection with repeated -fold cross-validation selects the best subset of features from eight predictors: highest price, lowest price, closing price, volume, change, price change ratio, and amplitude. These features are then used as input in a bidirectional gated recurrent unit with an attention mechanism (BiGRU-AM) model to forecast daily opening prices of ten stock indices. The proposed model exhibits superior forecasting performance across ten stock indices when compared to twelve benchmarks, evaluated using root mean squared error (RMSE), mean absolute error (MAE), and the coefficient of determination, . The improved prediction accuracy enables financial professionals to make more reliable investment decisions, reducing risks and increasing profits.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

Trading Signals (NASDAQ Composite Index Stock Forecast)

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

The heteroscedastic and volatile characteristics of stock price data have attracted the interest of researchers from various disciplines, particularly in the realm of price forecasting. The stock market’s non-stationary and volatile nature, driven by complex interrelationships among financial assets, economic developments, and market participants, poses significant challenges for accurate forecasting. This research aims to develop a robust forecasting model to improve the accuracy and reliability of stock price predictions using machine learning. A two-stage forecasting model is introduced. First, a random forest subset-based (RFS) feature selection with repeated -fold cross-validation selects the best subset of features from eight predictors: highest price, lowest price, closing price, volume, change, price change ratio, and amplitude. These features are then used as input in a bidirectional gated recurrent unit with an attention mechanism (BiGRU-AM) model to forecast daily opening prices of ten stock indices. The proposed model exhibits superior forecasting performance across ten stock indices when compared to twelve benchmarks, evaluated using root mean squared error (RMSE), mean absolute error (MAE), and the coefficient of determination, . The improved prediction accuracy enables financial professionals to make more reliable investment decisions, reducing risks and increasing profits.

In this paper, we introduce the mixed-frequency data model (MIDAS) to China’s insurance demand forecasting. We select the monthly indicators Consumer Confidence Index (CCI), China Economic Policy Uncertainty Index (EPU), Consumer Price Index (PPI), and quarterly indicator Depth of Insurance (TID) to construct a Mixed Data Sampling (MIDAS) regression model, which is used to study the impact and forecasting effect of CCI, EPU, and PPI on China’s insurance demand. To ensure forecasting accuracy, we investigate the forecasting effects of the MIDAS models with different weighting functions, forecasting windows, and a combination of forecasting methods, and use the selected optimal MIDAS models to forecast the short-term insurance demand in China. The experimental results show that the MIDAS model has good forecasting performance, especially in short-term forecasting. Rolling window and recursive identification prediction can improve the prediction accuracy, and the combination prediction makes the results more robust. Consumer confidence is the main factor influencing the demand for insurance during the COVID-19 period, and the demand for insurance is most sensitive to changes in consumer confidence. Shortly, China’s insurance demand is expected to return to the pre-COVID-19 level by 2023Q2, showing positive development. The findings of the study provide new ideas for China’s insurance policymaking.

I1 to I36 36 variables are selected as the inputs of the forecasting model. The name and description of the variables are shown in the 1st column and the 2nd column, respectively.Variables used as inputs.

2000-2010 Annual State-Scale Service and Domain scores used to support the approach for forecasting EPA's Human Well-Being Index. A modeling approach was developed based relationship function equations derived from select economic, social and ecosystem final goods and service scores and calculated human well-being index and related domain scores. These data are being used in a secondary capacity. The foundational data and scoring techniques were originally described in: a) U.S. EPA. 2012. Indicators and Methods for Constructing a U.S. Human Well-being Index (HWBI) for Ecosystem Services Research. Report. EPA/600/R-12/023. pp. 121; and b) U.S. EPA. 2014. Indicators and Methods for Evaluating Economic, Ecosystem and Social Services Provisioning. Report. EPA/600/R-14/184. pp. 174. Mode Smith, L. M., Harwell, L. C., Summers, J. K., Smith, H. M., Wade, C. M., Straub, K. R. and J.L. Case (2014).

This dataset is associated with the following publication: Summers , K., L. Harwell , and L. Smith. A Model For Change: An Approach for Forecasting Well-Being From Service-Based Decisions. ECOLOGICAL INDICATORS. Elsevier Science Ltd, New York, NY, USA, 69: 295-309, (2016).