Moving Average Research Articles

An In-Depth Look at Rising Temperatures: Forecasting with Advanced Time Series Models in Major US Regions

With growing concerns over climate change, accurately predicting temperature trends is crucial for informed decision-making and policy development. In this study, we perform a comprehensive comparative analysis of four advanced time series forecasting models—Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing (ETS), Multilayer Perceptron (MLP), and Gaussian Processes (GP)—to assess changes in minimum and maximum temperatures across four key regions in the United States. Our analysis includes hyperparameter optimization for each model to ensure peak performance. The results indicate that the MLP model outperforms the other models in terms of accuracy for temperature forecasting. Utilizing this best-performing model, we conduct temperature projections to evaluate the hypothesis that the rates of change in temperatures are greater than zero. Our findings confirm a positive rate of change in both maximum and minimum temperatures, suggesting a consistent upward trend over time. This research underscores the critical importance of refining time series forecasting models to address the challenges posed by climate change and supporting the development of effective strategies to mitigate the impacts of rising temperatures. The insights gained from this work emphasize the need for continuous advancement in predictive modeling techniques to better understand and respond to the dynamics of climate change.

ANALIZA STATYSTYK SIECI KOMPUTEROWYCH W CELU IDENTYFIKACJI PRZEPŁYWÓW INFORMACJI ZAKŁÓCAJĄCYCH STABILNOŚĆ W WOJSKOWYCH SIECIACH LOKALNYCH

The increasing complexity and scope of military computer networks necessitate robust methods to ensure network stability and security. This study presents a comprehensive analysis of computer network statistics in military local networks to develop a method for detecting information flows that disrupt stability. By leveraging advanced statistical techniques and machine learning algorithms, this research aims to enhance the cybersecurity posture of military local networks globally. Military networks are vital for communication, data exchange, and operational coordination. However, the dynamic nature of network traffic and the persistent threat of cyberattacks pose significant challenges to maintaining network stability. Traditional monitoring techniques often fail to meet the unique requirements of military networks, which demand high levels of security and rapid response capabilities. This study employs a multi-faceted approach to detect anomalies in network traffic, utilizing statistical methods such as Z-score analysis, Principal Component Analysis (PCA), and Autoregressive Integrated Moving Average (ARIMA) models. Machine learning techniques, including Support Vector Machines (SVM), Random Forests, Neural Networks, K-means clustering, and Reinforcement Learning, are also applied to identify patterns indicative of stability-disrupting information flows. The integration of statistical and machine learning methods forms a hybrid model that enhances anomaly detection, providing a robust framework for network security. The research problem is formulated as follows: does data collection include comprehensive network traffic data from various segments of military local area networks, including packet flows, transmission rates, and error rates over a specified period? Statistical analysis identifies patterns in the network traffic, which are then used to train machine learning models to classify normal and abnormal traffic. The research hypothesis states that machine learning models achieve high accuracy in detecting stability-disrupting information flows, with a precision rate exceeding 90%. The models identified several instances of stability-disrupting events, correlating these with known security incidents to validate the effectiveness of the detection method. This study underscores the importance of continuous monitoring and analysis of network statistics to ensure stability and security. The proposed method can be integrated with existing network monitoring and intrusion detection systems, providing a comprehensive approach to network security. Future research can build on these findings to develop more sophisticated models and explore additional factors influencing network stability, including the incorporation of advanced machine learning techniques, such as deep learning, and the exploration of other network metrics, like latency and packet loss. This comprehensive approach aims to enhance the security and operational reliability of military local networks.

Can customer engagement in social media be used as a predictor of operating revenue in SMEs?

Only a few studies have analyzed the direct or material contributions of customer engagement (CE) on social media. This is the first study that specifically examines the capability CE on social media to predict an indicator of business performance, such us the operating revenue, in a sample of small and medium-sized enterprises (SMEs). The author explores this issue by differentiating between the two main approaches used by SMEs in the management of their social networks (internal management with non-specialized staff and outsourced management with specialized professionals). The study, which uses time series analysis by applying the Autoregressive Integrated Moving Average (ARIMA) modeling technique, examines one hundred SMEs over an eight-year period (between October 2015 and October 2023). The results show that when social networks are managed internally, by non-specialized staff, CE on social media cannot be used as an indicator to forecast the company’s operating revenue. On the contrary, when the management of social networks is outsourced to specialized professionals (and combined with other digital marketing actions such as SEO, SEM, emailing, e-commerce, etc.), CE on social media can be used to predict the company's operating revenue. Consequently, based on the results obtained, the present study reveals that the role of CE on social media, as a driver of operating revenue, depends on the social network management approach adopted by the company on these platforms.

How Effective Are Forecasting Models in Predicting Effects of Exoskeletons on Fatigue Progression?

Forecasting can be utilized to predict future trends in physiological demands, which can be beneficial for developing effective interventions. This study implemented forecasting models to predict fatigue level progression when performing exoskeleton (EXO)-assisted tasks. Specifically, perceived and muscle activity data were utilized from nine recruited participants who performed 45° trunk flexion tasks intermittently with and without assistance until they reached medium-high exertion in the low-back region. Two forecasting algorithms, Autoregressive Integrated Moving Average (ARIMA) and Facebook Prophet, were implemented using perceived fatigue levels alone, and with external features of low-back muscle activity. Findings showed that univariate models without external features performed better with the Prophet model having the lowest mean (SD) of root mean squared error (RMSE) across participants of 0.62 (0.24) and 0.67 (0.29) with and without EXO-assisted tasks, respectively. Temporal effects of BSIE on delaying fatigue progression were then evaluated by forecasting back fatigue up to 20 trials. The slope of fatigue progression for 20 trials without assistance was ~48-52% higher vs. with assistance. Median benefits of 54% and 43% were observed for ARIMA (with external features) and Prophet algorithms, respectively. This study demonstrates some potential applications for forecasting models for workforce health monitoring, intervention assessment, and injury prevention.

Portfolio Evaluation of Selected Large cap & Midcap Mutual Fund Schemes

"Wealth, place & power are no measure of sources. The only true measure is the ratio between what we might have been & what we have become". In this regard people should upgrade their knowledge & keep in pace with changing technology & market condition to reach their desired goal. As new opportunities are opening up in every field all over the world. The desired goal can achieve only when both theory and practical go together hand in hand. This article is such an attempt to understand or link the theoretical concept with that of practical applicability. It provides an in-depth analysis of Mutual Funds Schemes, encompassing their benefits, various facts crucial to understand and associated risks. Furthermore, the theoretical background of mutual funds, including the types of funds available, their advantages, and potential risks have been dealt with. The study majorly focuses on evaluating the performance of specific mutual fund schemes in the large-cap and mid- cap categories. By analyzing various portfolio evaluation models, including Sharpe, Treynor, and Jensen's alpha, as well as other metrics such as Standard Deviation, Beta, and Portfolio Return, the research aims to identify top- performing funds. The study utilizes correlation analysis and the ARIMA model to forecast the Net Asset Value (NAV) of the highest-ranked funds, providing insights into risk and return profiles, and assisting advisors in recommending suitable investment choices to investors. The study concludes by emphasizing the importance of professional management and diversification in investment, particularly within the dynamic and unpredictable stock market environment. Keywords: Risk, Return, Beta, Standard Deviation (SD), Net Asset Value (NAV) & Assets under Management (AUM), Autoregressive Integrated Moving Average (ARIMA)

Trend Analysis of Tourist-arrivals in Nepal Using Auto- Regressive Integrated Moving Average (ARIMA) Model

Nepal, a diverse and geographically varied landlocked country with a rich tapestry of cultures, and natural heritages, boasts iconic attractions such as Mt. Everest as the world’s highest peak, Lumbini as the birthplace of Lord Buddha, Pashupatinath Temple as one of the world’s most celebrated pilgrimages for the Hindus, Chitwan National Park, home to the one-horned rhino enlisted by United Nations Educational, Scientific and Cultural Organization world’s natural treasure, etc. Attracting tourists from all corners of the globe, Nepal’s unique blend of landscapes and cultural heritage has made this country a prominent destination. This research delves into a comprehensive trend analysis of tourist arrivals in Nepal from 1964 to 2023, employing the Auto Regressive Integrated Moving Average (ARIMA) model based on the secondary data employed from the Ministry of Finance, Government of Nepal. The model provides insights into past trends and forecasts future tourist arrivals, evidenced by the rigorous test using the Root Mean Square Error. The analysis employed several ARIMA models, including ARIMA (2, 1, 3), ARIMA (2, 1, 2), ARIMA (3, 1, 3), and ARIMA (3, 1, 2), with model performance assessed using criteria such as Akaike Information Criterion (AIC) and Mean Absolute Error (MAE). Findings indicate that the ARIMA (2 1, 3) model provides the most accurate forecasts compared to the other models tested. This finding has significant implications for Nepal’s tourism industry. The results underscore the importance of adopting flexible and strategic planning approaches to enhance the resilience of Nepal’s tourism sector. The study’s insights are valuable for policymakers, industry stakeholders, and researchers, contributing to the development of targeted strategies that ensure sustainable growth in the face of ongoing volatility.

The impact of climate variation on the spread of Shiga toxin-producing E. coli.

Shiga toxin-producing Escherichia coli (STEC) is influenced by seasonality, but there is limited understanding of how specific climatic variables contribute to disease spread. This information aids in understanding disease transmission dynamics and could potentially inform public health modeling. This retrospective cohort study analyzed public health data from Ontario, Canada, between 2012 and 2021, along with historical climate data from Environment Canada. We employed Seasonal Autoregressive Integrated Moving Average (S-ARIMA) models to assess how temperature and precipitation impact the incidence of STEC infections, measured per 10,000,000 population. The study included 1658 confirmed STEC cases. A significant correlation was found between STEC incidence and climatic variables. Each degree Celsius increase in maximum temperature was associated with a rise of 3 STEC cases per 10,000,000 population (Centers for Disease Control and Prevention (2024)). Additionally, each millimeter of increased precipitation correlated with an increase of 1.1 cases per 10,000,000 population. The findings demonstrate a significant impact of temperature and precipitation on STEC transmission, highlighting the importance of integrating meteorological data into public health surveillance. This integration may help inform public health responses and support healthcare systems in planning for future outbreaks. Further studies are needed to refine predictive models and develop effective early warning systems for clinical settings.

Partial information sharing in supply chains with ARMA demand

AbstractIn this paper we suggest a novel mechanism for information sharing that allows a retailer to control the amount of shared information, and thus to limit information leakage, while still assisting the supplier to make better‐informed decisions and improve the overall efficiency of the supply chain. The control of the amount of leaked information facilitates information sharing because, absent such control, a retailer may refrain from sharing information due to the concern of information leakage. Specifically, we analyze a supply chain in which a retailer observes Autoregressive Moving Average (ARMA) demand for a single product where all players use the myopic order‐up‐to policy for determining their orders. We introduce a new class of information sharing arrangements, coined partial‐information shock (PaIS) sharing. This new class of information sharing agreements extends the previously studied mechanisms of demand sharing and full‐information shock sharing. We demonstrate that the retailer can construct a PaIS sharing arrangement that allows for an intermediate level of information sharing while simultaneously controlling the amount of leakage. We characterize when one PaIS arrangement will be more valuable to the supplier than another. We conclude with a numerical study that highlights that there does not necessarily need to be a tradeoff between the supplier having a better forecast and the retailer experiencing a higher level of leakage.

Trading Momentum in the U.S. Crude Oil Futures Market

This paper investigates if the Rate of Change (ROC), as a popular measure of momentum, can serve as a reliable technical analysis indicator to improve stock price prediction for U.S. West Texas Intermediate (WTI) crude oil futures market. The methodology centers on the application of the ROC and/or Moving Average (MA) price crossover/cross under strategies as a trading system. End of month futures prices of West Texas Intermediate (WTI) crude oil prices are collected for the period 28th May 2004 – 30th April 2024. The performance of the trading system is measured using both the Sharpe and Sortino ratios, thereby adjusting for total and downside risks. The model is also benchmarked against the naïve buy-and-hold strategy. Overall findings suggest a ROC based on 14-month periods outperform other momentum-based indicators, including when combined with price-crossover moving average strategies, and the naïve buy-and-hold strategy. After adjusting for the negative returns, the downside risk or semi-deviation amounted to 8.5%, and a Sortino value of 4.58. The Sortino value is however biased due to the 295% return witnessed in 2009. Findings have some vital policy implications for regulatory bodies and traders in the WTI crude oil energy futures market.

Groundwater table prediction and seasonal variation influenced by short rotation willow plantation on marginal riparian lands of the Prairie potholes in Canada

Shallow groundwater consumption via phreatophytic transpiration and resulting vegetation-linked groundwater table (GWT) fluctuation is a typical soil hydrological process in wetland riparian areas. However, upland and riparian land use alterations may further influence the shallow GWT fluctuation, temporally and spatially. In this multi-year field study, we investigated whether introducing short rotation willow (SRW) positively or negatively affects the shallow GWT, soil water availability, and soil health on marginal riparian lands of the Prairie Pothole Region (PPR). We compared the impact of SRW on these parameters to two common land uses: annual crop (AC) and pasture (PA). Depth to GWT was monitored via data loggers from 28 wells in two semi-arid PPR sites. The GWT depth varied by land use practices only in site B (p < 0.001; PA > SRW = AC) but not significantly in site A (p = 0.325), and the patterns were inconsistent between sites. In GWT depth prediction, the performance of Artificial Neural Network (ANN) was better than Autoregressive Integrated Moving Average (ARIMA) models but was inconsistent alike with field observations. The GWT depth responded to seasonal precipitation and potential evapotranspiration (ET) patterns. The monthly GWT fluctuations peaked between June and August due to increased precipitation, while they were lower during May and September with reduced precipitation; however, these variations were not significant (p > 0.05). Higher precipitation and lower potential ET throughout the wet year (i.e., in 2014) significantly (p < 0.05) raised GWT (i.e., decreased depth to GWT) under all land uses, and vice versa. Our study indicated that planting SRW in marginal riparian land of the PPR would not negatively impact shallow GWT or soil water availability. Moreover, the SRW plantation could also help manage soil salinity without severely depleting the soil's nutrient pools or diminishing soil quality and health indicator parameters measured during the first rotation.

Analyzing the Selective Stock Price Index Using Fractionally Integrated and Heteroskedastic Models

Stock market indices are important tools to measure and compare stock market performance. The Selective Stock Price (SSP) index reflects fluctuations in a set value of financial instruments of Santiago de Chile’s stock exchange. Stock indices also reflect volatility linked to high uncertainty or potential investment risk. However, economic shocks are altering volatility. Evidence of long memory in SSP time series also exists, which implies long-term persistence. In this paper, we studied the volatility of SSP time series from January 2010 to September 2023 using fractionally heteroskedastic models. We considered the Autoregressive Fractionally Integrated Moving Average (ARFIMA) process with Generalized Autoregressive Conditional Heteroskedasticity (GARCH) innovations—the ARFIMA-GARCH model—for SSP log returns, and the fractionally integrated GARCH, or FIGARCH model, was compared with a classical GARCH one. The results show that the ARFIMA-GARCH model performs best in terms of volatility fit and predictive quality. This model allows us to obtain a better understanding of the observed volatility and its behavior, which contributes to more effective investment risk management in the stock market. Moreover, the proposed model detects the influence volatility increments of the SSP index linked to external factors that impact the economic outlook, such as China’s economic slowdown in 2012 and the subprime crisis in 2008.

Patterns in the Chaos: The Moving Hurst Indicator and Its Role in Indian Market Volatility

Estimating the impact of volatility in financial markets is challenging due to complex dynamics, including random fluctuations involving white noise and trend components involving brown noise. In this study, we explore the potential of leveraging the chaotic properties of time series data for improved accuracy. Specifically, we introduce a novel trading strategy based on a technical indicator, Moving Hurst (MH). MH utilizes the Hurst exponent which characterizes the chaotic properties of time series. We hypothesize and then prove empirically that MH outperforms traditional indicators like Moving Averages (MA) in analyzing Indian equity indices and capturing profitable trading opportunities while mitigating the impact of volatility.

Prediction of ionospheric TEC during the occurrence of earthquakes in Indonesia using ARMA and CoK models

Predicting ionospheric Total Electron Content (TEC) variations associated with seismic activity is crucial for mitigating potential disruptions in communication networks, particularly during earthquakes. This research investigates applying two modelling techniques, Autoregressive Moving Average (ARMA) and Cokriging (CoK) based models to forecast ionospheric TEC changes linked to seismic events in Indonesia. The study focuses on two significant earthquakes: the December 2004 Sumatra earthquake and the August 2012 Sulawesi earthquake. GPS TEC data from a BAKO station near Indonesia and solar and geomagnetic data were utilized to assess the causes of TEC variations. The December 2004 Sumatra earthquake, registering a magnitude of 9.1–9.3, exhibited notable TEC variations 5 days before the event. Analysis revealed that the TEC variations were weakly linked to solar and geomagnetic activities. Both ARMA and CoK models were employed to predict TEC variations during the Earthquakes. The ARMA model demonstrated a maximum TEC prediction of 50.92 TECU and a Root Mean Square Error (RMSE) value of 6.15, while the CoK model predicted a maximum TEC of 50.68 TECU with an RMSE value of 6.14. The August 2012 Sulawesi earthquake having a magnitude of 6.6, revealed TEC anomalies 6 days before the event. For both the Sumatra and Sulawesi earthquakes, the GPS TEC variations showed weak associations with solar and geomagnetic activities but stronger correlations with the earthquake-induced electric field for the considered two stations. The ARMA model predicted a maximum TEC of 54.43 TECU with an RMSE of 3.05, while the CoK model predicted a maximum TEC of 52.90 TECU with an RMSE of 7.35. Evaluation metrics including RMSE, Mean Absolute Deviation (MAD), Relative Error, and Normalized RMSE (NRMSE) were employed to assess the accuracy and reliability of the prediction models. The results indicated that while both models captured the general trend in TEC variations, nuances emerged in their responses to seismic events. The ARMA model demonstrated heightened sensitivity to seismic disturbances, particularly evident on the day of the earthquake, whereas the CoK model exhibited more consistent performance across pre- and post-earthquake periods.

CLARA: A cluster-based node correlation for sampling rate adaptation and fault tolerance in sensor networks

Recently, wireless sensor networks (WSNs) have been proven as an efficient and low-cost solution for monitoring various kind of applications. However, the massive amount of data collected and transmitted by the sensor nodes, which are mostly redundant, will quickly consume their limited battery power, which is sometimes difficult to replace or recharge. Although the huge efforts made by researchers to solve such problem, most of the proposed techniques suffer from their accuracy and their complexity, which is not suitable for limited-resources sensors. Therefore, designing new data reduction techniques to reduce the raw data collected in such networks is becoming essential to increase their lifetime. In this paper, we propose a CLuster-based node correlation for sAmpling Rate adaptation and fAult tolerance, abbreviated CLARA, mechanism dedicated to periodic sensor network applications. Mainly, CLARA works on two stages: node correlation and fault tolerance. The first stage introduces a data clustering method that aims to search the correlation among neighboring nodes. Then, it accordingly adapts their sensing frequencies in a way to reduce the amount of data collected in such networks while preserving the information integrity at the sink. In the second stage, a fault tolerance model is proposed that allows the sink to regenerate the raw sensor data based on two methods: moving average (MA) and exponential smoothing (ES). We demonstrated the efficiency of our technique through both simulations and experiments. The best obtained results show that the first stage can reduce the sensor sampling rate, and accordingly the sensor energy, up to 64% while the second stage can accurately regenerate the raw data with an error loss less than 0.15.

Contextual Intelligence: An AI Approach to Manufacturing Skills’ Forecasting

The manufacturing industry is skill-intensive and plays a pivotal role in South Africa’s economy, reflecting the nation’s progress and development. The advent of technology has initiated a transformative era within the manufacturing sector. Workforce skills are at the heart of ensuring the sustained growth of the industry. This study delves into the skill-related aspects of the occupational landscape of the South African manufacturing sector, with a particular focus on two important manufacturing sectors: the food and beverage manufacturing (FoodBev) sector and the chemical manufacturing (CHIETA) sector. Leveraging the forecasting prowess of Autoregressive Integrated Moving Average (ARIMA), this paper outlines a sectorial occupational forecasting modeling exercise to reveal which job roles are poised for expansion and which are expected to decline. The approach predicted future skills’ demand 80% accuracy for 473 out of 713 (66%) occupations for FoodBev and 474 out of 522 (91%) for CHIETA. These insights are invaluable for industry stakeholders and educational institutions, providing guidance to support the sector’s growth in an era marked by technological advancement.

Carbon Emission Prediction Algorithm for Community Electricity Consumption Based on Improved PSO Algorithm with Optimized Autoregressive Moving Average Models

Abstract The anticipation of carbon emissions stemming from community electricity usage stands as a pivotal field of study. Such analysis holds the potential to furnish communities with informed pathways towards crafting viable carbon reduction strategies, thereby contributing significantly to the overarching goals of carbon peaking and eventual neutrality. The intricacies of carbon emissions from community electricity consumption are manifold, entangled with variables like energy consumption patterns, energy mix, and carbon emission coefficients. Hence, the imperative lies in crafting a predictive framework adept at holistically integrating these variables to enhance prognostic precision and reliability. Within this research, we propose a dynamic carbon emission factor regression model. This model is uniquely poised to capture real-time shifts influenced by changes in energy structure and policy landscapes, thereby amplifying predictive sensitivity. Leveraging the Particle Swarm Optimization (PSO) algorithm, we synchronously optimize the autoregressive terms (p) of the Autoregressive Integrated Moving Average (ARIMA) model and the moving average terms (q) of the MA model to attain a globally optimal solution. Crucially, this approach obviates the need for manual intervention and arbitrary parameter selection. In contrast to conventional optimization methodologies, our paper advances the PSO’s weight calculation mechanism. By assigning greater weight values during the initial iterations, the algorithm maintains robust global search capabilities. Subsequently, the inertia weight diminishes progressively throughout the iteration process, fostering precise local exploration.

Advancing predictive accuracy of shallow landslide using strategic data augmentation

Rainfall-induced shallow landslides pose one of significant geological hazards, necessitating precise monitoring and prediction for effective disaster mitigation. Most studies on landslide prediction have focused on optimizing machine learning (ML) algorithms, very limited attention has been paid to enhancing data quality for improved predictive performance. This study employs strategic data augmentation (DA) techniques to enhance the accuracy of shallow landslide prediction. Using five DA methods including singular spectrum analysis (SSA), moving averages (MA), wavelet denoising (WD), variational mode decomposition (VMD), and linear interpolation (LI), we utilize strategies such as smoothing, denoising, trend decomposition, and synthetic data generation to improve the training dataset. Four machine learning algorithms, i.e. artificial neural network (ANN), recurrent neural network (RNN), one-dimensional convolutional neural network (CNN1D), and long short-term memory (LSTM), are used to forecast landslide displacement. The case study of a landslide in southwest China shows the effectiveness of our approach in predicting landslide displacements, despite the inherent limitations of the monitoring dataset. VMD proves the most effective for smoothing and denoising, improving R2, RMSE, and MAPE by 172.16%, 71.82%, and 98.9%, respectively. SSA addresses missing data, while LI is effective with limited data samples, improving metrics by 21.6%, 52.59%, and 47.87%, respectively. This study demonstrates the potential of DA techniques to mitigate the impact of data defects on landslide prediction accuracy, with implications for similar cases.

Focus-SLAM: A visual Monocular SLAM

Abstract This research presents a novel simultaneous localization and mapping algorithm, called Focus-SLAM, which simulates human navigation strategies by synthesizing a optical saliency model, SalNavNet, in a traditional single-shot SLAM paradigm. SalNavNet introduces an innovative design that incorporates a correlation module and an adaptive Exponential Moving Average (EMA) component, effectively addressing the prevalent center bias issue found in contemporary saliency models. As a result, the system enhances target fixation by accentuating salient features more effectively. Extensive experimental evaluations, conducted across a spectrum of indoor environments exhibiting diverse luminosity conditions. These findings attest to the efficacy of Focus-SLAM in enhancing both the accuracy and efficiency of SLAM operations under challenging real-world conditions.

Fuzzified input data tuning for agriculture commodities price prediction

&lt;span lang="EN-US"&gt;The quality of the input data will typically affect the prediction accuracy. Preprocessing of data is commonly referred to as input data tuning. Tuning the input data is critical for projecting commodity prices. Anomalies or outliers are unavoidable in historical price data. To increase prediction and forecasting accuracy, it is necessary to find and correct outliers before training the prediction model. To correct the anomaly and increase prediction accuracy, the Fuzzified Input Data Tuning and Prediction algorithm proposed in this study. The identified outliers are corrected using the relevant fuzzy set value in this method. With outlier corrected data, we used Long Short-Term Memory and Seasonal Autoregressive Integrated Moving Average to anticipate tomato prices in Karnataka state. The result of the proposed algorithm is compared with the Sliding Window anomalies correction model, and without disposing of the outliers. The suggested algorithm, with 37.89%, performed better than Sliding Window with 40.08% and 43.11% Mean Absolute Percentage Error, respectively, and without outlier correction. The sensitivity analysis shows that the performance of the model is unaffected by the forecasting horizon. Finally, comparitive analysis peformed with previous research work, and the proposed model performed better.&lt;/span&gt;

Foreign Exchange Rates Prediction Based on Comparative Models

With the development of globalization, fluctuations in the foreign exchange market have the huge potential to influence global economy. Machine Learning technology shows great potential in the field of financial forecasting than traditional statistical methods in recent years. This research selected statistical model Autoregressive Integrated Moving Average (ARIMA) and machine learning models Linear Regression and Long- and Short-Term Memory (LSTM) to make training and testing of historical data on the forex market. Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-Squared were chosen as evaluation indexes to evaluate the performance of each model and compare their performance in forex exchange rate forecasts. The results show that ARIMA and LSTM could make better predictions than Linear Regression, between which LSTM has higher degree of fitting data. This research provides a detailed analysis comparing the performance of different models in forex exchange rate forecasting whose results are beneficial to guide forex market forecast in practical application and provide reference for future research.