Search results for: out-of-sample forecasts

Authors: Vikrant Gupta, Amrit Goswami

Abstract:

The fixed income market forms the basis of the modern financial market. All other assets in financial markets derive their value from the bond market. Owing to its over-the-counter nature, corporate bonds have relatively less data publicly available and thus is researched upon far less compared to Equities. Bond price prediction is a complex financial time series forecasting problem and is considered very crucial in the domain of finance. The bond prices are highly volatile and full of noise which makes it very difficult for traditional statistical time-series models to capture the complexity in series patterns which leads to inefficient forecasts. To overcome the inefficiencies of statistical models, various machine learning techniques were initially used in the literature for more accurate forecasting of time-series. However, simple machine learning methods such as linear regression, support vectors, random forests fail to provide efficient results when tested on highly complex sequences such as stock prices and bond prices. hence to capture these intricate sequence patterns, various deep learning-based methodologies have been discussed in the literature. In this study, a recurrent neural network-based deep learning model using long short term networks for prediction of corporate bond prices has been discussed. Long Short Term networks (LSTM) have been widely used in the literature for various sequence learning tasks in various domains such as machine translation, speech recognition, etc. In recent years, various studies have discussed the effectiveness of LSTMs in forecasting complex time-series sequences and have shown promising results when compared to other methodologies. LSTMs are a special kind of recurrent neural networks which are capable of learning long term dependencies due to its memory function which traditional neural networks fail to capture. In this study, a simple LSTM, Stacked LSTM and a Masked LSTM based model has been discussed with respect to varying input sequences (three days, seven days and 14 days). In order to facilitate faster learning and to gradually decompose the complexity of bond price sequence, an Empirical Mode Decomposition (EMD) has been used, which has resulted in accuracy improvement of the standalone LSTM model. With a variety of Technical Indicators and EMD decomposed time series, Masked LSTM outperformed the other two counterparts in terms of prediction accuracy. To benchmark the proposed model, the results have been compared with traditional time series models (ARIMA), shallow neural networks and above discussed three different LSTM models. In summary, our results show that the use of LSTM models provide more accurate results and should be explored more within the asset management industry.

Keywords: bond prices, long short-term memory, time series forecasting, empirical mode decomposition

Procedia PDF Downloads 115

Authors: Egemen Sert, Leila Hedayatifar, Rachel A. Rigg, Amir Akhavan, Olha Buchel, Dominic Elias Saadi, Aabir Abubaker Kar, Alfredo J. Morales, Yaneer Bar-Yam

Abstract:

The complexity of providing timely and cost-effective distribution of finished goods from industrial facilities to customers makes effective operational coordination difficult, yet effectiveness is crucial for maintaining customer service levels and sustaining a business. Logistics planning becomes increasingly complex with growing numbers of customers, varied geographical locations, the uncertainty of future orders, and sometimes extreme competitive pressure to reduce inventory costs. Linear optimization methods become cumbersome or intractable due to a large number of variables and nonlinear dependencies involved. Here we develop a complex systems approach to optimizing logistics networks based upon dimensional reduction methods and apply our approach to a case study of a manufacturing company. In order to characterize the complexity in customer behavior, we define a “customer space” in which individual customer behavior is described by only the two most relevant dimensions: the distance to production facilities over current transportation routes and the customer's demand frequency. These dimensions provide essential insight into the domain of effective strategies for customers; direct and indirect strategies. In the direct strategy, goods are sent to the customer directly from a production facility using box or bulk trucks. In the indirect strategy, in advance of an order by the customer, goods are shipped to an external warehouse near a customer using trains and then "last-mile" shipped by trucks when orders are placed. Each strategy applies to an area of the customer space with an indeterminate boundary between them. Specific company policies determine the location of the boundary generally. We then identify the optimal delivery strategy for each customer by constructing a detailed model of costs of transportation and temporary storage in a set of specified external warehouses. Customer spaces help give an aggregate view of customer behaviors and characteristics. They allow policymakers to compare customers and develop strategies based on the aggregate behavior of the system as a whole. In addition to optimization over existing facilities, using customer logistics and the k-means algorithm, we propose additional warehouse locations. We apply these methods to a medium-sized American manufacturing company with a particular logistics network, consisting of multiple production facilities, external warehouses, and customers along with three types of shipment methods (box truck, bulk truck and train). For the case study, our method forecasts 10.5% savings on yearly transportation costs and an additional 4.6% savings with three new warehouses.

Keywords: logistics network optimization, direct and indirect strategies, K-means algorithm, dimensional reduction

Procedia PDF Downloads 118

Authors: Monika Saha, A. T. M. Hasan Zobeyer, Nasreen Jahan

Abstract:

El Nino-Southern Oscillation (ENSO) is the interaction between atmosphere and ocean in tropical Pacific which causes inconsistent warm/cold weather in tropical central and eastern Pacific Ocean. Due to the impact of climate change, ENSO events are becoming stronger in recent times, and therefore it is very important to study the influence of ENSO in climate studies. Bangladesh, being in the low-lying deltaic floodplain, experiences the worst consequences due to flooding every year. To reduce the catastrophe of severe flooding events, non-structural measures such as flood forecasting can be helpful in taking adequate precautions and steps. Forecasting seasonal flood with a longer lead time of several months is a key component of flood damage control and water management. The objective of this research is to identify the possible strength of teleconnection between ENSO and river flow of Surma and examine the potential possibility of long lead flood forecasting in the wet season. Surma is one of the major rivers of Bangladesh and is a part of the Surma-Meghna river system. In this research, sea surface temperature (SST) has been considered as the ENSO index and the lead time is at least a few months which is greater than the basin response time. The teleconnection has been assessed by the correlation analysis between July-August-September (JAS) flow of Surma and SST of Nino 4 region of the corresponding months. Cumulative frequency distribution of standardized JAS flow of Surma has also been determined as part of assessing the possible teleconnection. Discharge data of Surma river from 1975 to 2015 is used in this analysis, and remarkable increased value of correlation coefficient between flow and ENSO has been observed from 1985. From the cumulative frequency distribution of the standardized JAS flow, it has been marked that in any year the JAS flow has approximately 50% probability of exceeding the long-term average JAS flow. During El Nino year (warm episode of ENSO) this probability of exceedance drops to 23% and while in La Nina year (cold episode of ENSO) it increases to 78%. Discriminant analysis which is known as 'Categoric Prediction' has been performed to identify the possibilities of long lead flood forecasting. It has helped to categorize the flow data (high, average and low) based on the classification of predicted SST (warm, normal and cold). From the discriminant analysis, it has been found that for Surma river, the probability of a high flood in the cold period is 75% and the probability of a low flood in the warm period is 33%. A synoptic parameter, forecasting index (FI) has also been calculated here to judge the forecast skill and to compare different forecasts. This study will help the concerned authorities and the stakeholders to take long-term water resources decisions and formulate policies on river basin management which will reduce possible damage of life, agriculture, and property.

Keywords: El Nino-Southern Oscillation, sea surface temperature, surma river, teleconnection, cumulative frequency distribution, discriminant analysis, forecasting index

Procedia PDF Downloads 126

Authors: Tomola Obamuyi

Abstract:

The banking system in Nigeria adopted innovative banking, with the aim of enhancing financial inclusion, and making financial services readily and cheaply available to majority of the people, and to contribute to the efficiency of the financial system. Some of the innovative services include: Automatic Teller Machines (ATMs), National Electronic Fund Transfer (NEFT), Point of Sale (PoS), internet (Web) banking, Mobile Money payment (MMO), Real-Time Gross Settlement (RTGS), agent banking, among others. The introduction of these payment systems is expected to increase bank efficiency and customers' satisfaction, culminating in better performance for the commercial banks. However, opinions differ on the possible effects of the various innovative payment systems on the performance of commercial banks in the country. Thus, this study empirically determines how commercial banks use innovation to gain competitive advantage in the specific context of Nigeria's finance and business. The study also analyses the effects of financial innovation on the performance of commercial banks, when different periods of analysis are considered. The study employed secondary data from 2009 to 2018, the period that witnessed aggressive innovation in the financial sector of the country. The Vector Autoregression (VAR) estimation technique forecasts the relative variance of each random innovation to the variables in the VAR, examine the effect of standard deviation shock to one of the innovations on current and future values of the impulse response and determine the causal relationship between the variables (VAR granger causality test). The study also employed the Multi-Criteria Decision Making (MCDM) to rank the innovations and the performance criteria of Return on Assets (ROA) and Return on Equity (ROE). The entropy method of MCDM was used to determine which of the performance criteria better reflect the contributions of the various innovations in the banking sector. On the other hand, the Range of Values (ROV) method was used to rank the contributions of the seven innovations to performance. The analysis was done based on medium term (five years) and long run (ten years) of innovations in the sector. The impulse response function derived from the VAR system indicated that the response of ROA to the values of cheques transaction, values of NEFT transactions, values of POS transactions was positive and significant in the periods of analysis. The paper also confirmed with entropy and range of value that, in the long run, both the CHEQUE and MMO performed best while NEFT was next in performance. The paper concluded that commercial banks would enhance their performance by continuously improving on the services provided through Cheques, National Electronic Fund Transfer and Point of Sale since these instruments have long run effects on their performance. This will increase the confidence of the populace and encourage more usage/patronage of these services. The banking sector will in turn experience better performance which will improve the economy of the country. Keywords: Bank performance, financial innovation, multi-criteria decision making, vector autoregression,

Keywords: Bank performance, financial innovation, multi-criteria decision making, vector autoregression

Procedia PDF Downloads 94

Authors: Piotr A. Kowalski, Kasper Sapala, Wiktor Warchalowski

Abstract:

PM10 is a suspended dust that primarily has a negative effect on the respiratory system. PM10 is responsible for attacks of coughing and wheezing, asthma or acute, violent bronchitis. Indirectly, PM10 also negatively affects the rest of the body, including increasing the risk of heart attack and stroke. Unfortunately, Poland is a country that cannot boast of good air quality, in particular, due to large PM concentration levels. Therefore, based on the dense network of Airly sensors, it was decided to deal with the problem of prediction of suspended particulate matter concentration. Due to the very complicated nature of this issue, the Machine Learning approach was used. For this purpose, Convolution Neural Network (CNN) neural networks have been adopted, these currently being the leading information processing methods in the field of computational intelligence. The aim of this research is to show the influence of particular CNN network parameters on the quality of the obtained forecast. The forecast itself is made on the basis of parameters measured by Airly sensors and is carried out for the subsequent day, hour after hour. The evaluation of learning process for the investigated models was mostly based upon the mean square error criterion; however, during the model validation, a number of other methods of quantitative evaluation were taken into account. The presented model of pollution prediction has been verified by way of real weather and air pollution data taken from the Airly sensor network. The dense and distributed network of Airly measurement devices enables access to current and archival data on air pollution, temperature, suspended particulate matter PM1.0, PM2.5, and PM10, CAQI levels, as well as atmospheric pressure and air humidity. In this investigation, PM2.5, and PM10, temperature and wind information, as well as external forecasts of temperature and wind for next 24h served as inputted data. Due to the specificity of the CNN type network, this data is transformed into tensors and then processed. This network consists of an input layer, an output layer, and many hidden layers. In the hidden layers, convolutional and pooling operations are performed. The output of this system is a vector containing 24 elements that contain prediction of PM10 concentration for the upcoming 24 hour period. Over 1000 models based on CNN methodology were tested during the study. During the research, several were selected out that give the best results, and then a comparison was made with the other models based on linear regression. The numerical tests carried out fully confirmed the positive properties of the presented method. These were carried out using real ‘big’ data. Models based on the CNN technique allow prediction of PM10 dust concentration with a much smaller mean square error than currently used methods based on linear regression. What's more, the use of neural networks increased Pearson's correlation coefficient (R²) by about 5 percent compared to the linear model. During the simulation, the R² coefficient was 0.92, 0.76, 0.75, 0.73, and 0.73 for 1st, 6th, 12th, 18th, and 24th hour of prediction respectively.

Keywords: air pollution prediction (forecasting), machine learning, regression task, convolution neural networks

Procedia PDF Downloads 122

Authors: K. Bencherif, M. Bellifa

Abstract:

The expected climatic changes in North Africa are the increase of both intensity and frequencies of the summer droughts and a reduction in water availability during growing season. The exiting coppices and forest formations in the national park of Tlemcen are dominated by holm oak, zen oak and cork oak. These opened-fragmented structures don’t seem enough strong so to hope durable protection against climate change. According to the observed climatic tendency, the objective is to analyze the climatic context and its evolution taking into account the eventual behaving of the oak species during the next 20-30 years on one side and the landscaped context in relation with the most adequate sylvicultural models to choose and especially in relation with human activities on another side. The study methodology is based on Climatic synthesis and Floristic and spatial analysis. Meteorological data of the decade 1989-2009 are used to characterize the current climate. An another approach, based on dendrochronological analysis of a 120 years sample Aleppo pine stem growing in the park, is used so to analyze the climate evolution during one century. Results on the climate evolution during the 50 years obtained through climatic predictive models are exploited so to predict the climate tendency in the park. Spatially, in each forest unit of the Park, stratified sampling is achieved so to reduce the degree of heterogeneity and to easily delineate different stands using the GPS. Results from precedent study are used to analyze the anthropogenic factor considering the forecasts for the period 2025-2100, the number of warm days with a temperature over 25°C would increase from 30 to 70. The monthly mean temperatures of the maxima’s (M) and the minima’s (m) would pass respectively from 30.5°C to 33°C and from 2.3°C to 4.8°C. With an average drop of 25%, precipitations will be reduced to 411.37 mm. These new data highlight the importance of the risk fire and the water stress witch would affect the vegetation and the regeneration process. Spatial analysis highlights the forest and the agricultural dimensions of the park compared to the urban habitat and bare soils. Maps show both fragmentation state and forest surface regression (50% of total surface). At the level of the park, fires affected already all types of covers creating low structures with various densities. On the silvi cultural plan, Zen oak form in some places pure stands and this invasion must be considered as a natural tendency where Zen oak becomes the structuring specie. Climate-related changes have nothing to do with the real impact that South-Mediterranean forests are undergoing because human constraints they support. Nevertheless, hardwoods stand of oak in the national park of Tlemcen will face up to unexpected climate changes such as changing rainfall regime associated with a lengthening of the period of water stress, to heavy rainfall and/or to sudden cold snaps. Faced with these new conditions, management based on mixed uneven aged high forest method promoting the more dynamic specie could be an appropriate measure.

Keywords: global warming, mediterranean forest, oak shrub-lands, Tlemcen

Procedia PDF Downloads 373

Authors: Sabrina Neururer, Marco Schweitzer, Werner Hackl, Bernhard Tilg, Patrick Raudaschl, Andreas Huber, Bernhard Pfeifer

Abstract:

History and the current outbreak of Covid-19 have shown the deadly potential of infectious diseases. However, infectious diseases also have a serious impact on areas other than health and healthcare, such as the economy or social life. These areas are strongly codependent. Therefore, disease control measures, such as social distancing, quarantines, curfews, or lockdowns, have to be adopted in a very considerate manner. Infectious disease modeling can support policy and decision-makers with adequate information regarding the dynamics of the pandemic and therefore assist in planning and enforcing appropriate measures that will prevent the healthcare system from collapsing. In this work, an agent-based simulation package named “survival” for simulating infectious diseases is presented. A special focus is put on SARS-Cov-2. The presented simulation package was used in Austria to model the SARS-Cov-2 outbreak from the beginning of 2020. Agent-based modeling is a relatively recent modeling approach. Since our world is getting more and more complex, the complexity of the underlying systems is also increasing. The development of tools and frameworks and increasing computational power advance the application of agent-based models. For parametrizing the presented model, different data sources, such as known infections, wastewater virus load, blood donor antibodies, circulating virus variants and the used capacity for hospitalization, as well as the availability of medical materials like ventilators, were integrated with a database system and used. The simulation result of the model was used for predicting the dynamics and the possible outcomes and was used by the health authorities to decide on the measures to be taken in order to control the pandemic situation. The survival package was implemented in the programming language Java and the analytics were performed with R Studio. During the first run in March 2020, the simulation showed that without measures other than individual personal behavior and appropriate medication, the death toll would have been about 27 million people worldwide within the first year. The model predicted the hospitalization rates (standard and intensive care) for Tyrol and South Tyrol with an accuracy of about 1.5% average error. They were calculated to provide 10-days forecasts. The state government and the hospitals were provided with the 10-days models to support their decision-making. This ensured that standard care was maintained for as long as possible without restrictions. Furthermore, various measures were estimated and thereafter enforced. Among other things, communities were quarantined based on the calculations while, in accordance with the calculations, the curfews for the entire population were reduced. With this framework, which is used in the national crisis team of the Austrian province of Tyrol, a very accurate model could be created on the federal state level as well as on the district and municipal level, which was able to provide decision-makers with a solid information basis. This framework can be transferred to various infectious diseases and thus can be used as a basis for future monitoring.

Keywords: modelling, simulation, agent-based, SARS-Cov-2, COVID-19

Procedia PDF Downloads 155

Authors: Adrian Priceputu, Elena Mihaela Stan

Abstract:

Our research addresses a critical challenge in renewable energy: improving efficiency and reducing the costs associated with the installation of ground-mounted photovoltaic (PV) panels. The most commonly used foundation solution is metal piles - with various sections adapted to soil conditions and the structural model of the panels. However, direct foundation systems are also sometimes used, especially in brownfield sites. Although metal micropiles are generally the first design option, understanding and predicting their bearing capacity, particularly under varied soil conditions, remains an open research topic. CPT Method and Current Challenges: Metal piles are favored for PV panel foundations due to their adaptability, but existing design methods rely heavily on costly and time-consuming in situ tests. The Cone Penetration Test (CPT) offers a more efficient alternative by providing valuable data on soil strength, stratification, and other key characteristics with reduced resources. During the test, a cone-shaped probe is pushed into the ground at a constant rate. Sensors within the probe measure the resistance of the soil to penetration, divided into cone penetration resistance and shaft friction resistance. Despite some existing CPT-based design approaches for metal piles, these methods are often cumbersome and difficult to apply. They vary significantly due to soil type and foundation method, and traditional approaches like the LCPC method involve complex calculations and extensive empirical data. The method was developed by testing 197 piles on a wide range of ground conditions, but the tested piles were very different from the ones used for PV pile foundations, making the method less accurate and practical for steel micropiles. Project Objectives and Methodology: Our research aims to develop a calculation method for metal micropile foundations using CPT data, simplifying the complex relationships involved. The goal is to estimate the pullout bearing capacity of piles without additional laboratory tests, streamlining the design process. To achieve this, a case study was selected which will serve for the development of an 80ha solar power station. Four testing locations were chosen spread throughout the site. At each location, two types of steel profiles (H160 and C100) were embedded into the ground at various depths (1.5m and 2.0m). The piles were tested for pullout capacity under natural and inundated soil conditions. CPT tests conducted nearby served as calibration points. The results served for the development of a preliminary equation for estimating pullout capacity. Future Work: The next phase involves validating and refining the proposed equation on additional sites by comparing CPT-based forecasts with in situ pullout tests. This validation will enhance the accuracy and reliability of the method, potentially transforming the foundation design process for PV panels.

Keywords: cone penetration test, foundation optimization, solar power stations, steel pile foundations

Procedia PDF Downloads 18

Authors: Caridad Maylin-Aguilar, Beatriz Duarte-Monedero

Abstract:

Problem-based learning (PBL) is a useful tool for adult and practice oriented audiences, as University students. As a consequence of the huge disruption caused by the COVID pandemic in the hospitality industry, hotels of all categories closed down in Spain from March 2020. Since that moment, the luxury segment was blooming with optimistic prospects for new openings. Hence, Hospitality students were expecting a positive situation in terms of employment and career development. By the beginning of the 2020-21 academic year, these expectations were seriously harmed. By October 2020, only 9 of the 32 hotels in the luxury segment were opened with an occupation rate of 9%. Shortly after, the evidence of a second wave affecting especially Spain and the homelands of incoming visitors bitterly smashed all forecasts. In accordance with the situation, a team of four professors and practitioners, from four different subject areas, developed a real case, inspired in one of these hotels, the 5-stars Emperatriz by Barceló. Students in their 2nd course were provided with real information as marketing plans, profit and losses and operational accounts, employees profiles and employment costs. The challenge for them was to act as consultants, identifying potential courses of action, related to best, base and worst case. In order to do that, they were organized in teams and supported by 4th course students. Each professor deployed the problem in their subject; thus, research on the customers behavior and feelings were necessary to review, as part of the marketing plan, if the current offering of the hotel was clear enough to guarantee and to communicate a safe environment, as well as the ranking of other basic, supporting and facilitating services. Also, continuous monitoring of competitors’ activity was necessary to understand what was the behavior of the open outlets. The actions designed after the diagnose were ranked in accordance with their impact and feasibility in terms of time and resources. Also they must be actionable by the current staff of the hotel and their managers and a vision of internal marketing was appreciated. After a process of refinement, seven teams presented their conclusions to Emperatriz general manager and the rest of professors. Four main ideas were chosen, and all the teams, irrespectively of authorship, were asked to develop them to the state of a minimum viable product, with estimations of impacts and costs. As the process continues, students are nowadays accompanying the hotel and their staff in the prudent reopening of facilities, almost one year after the closure. From a professor’s point of view, key learnings were 1.- When facing a real problem, a holistic view is needed. Therefore, the vision of subjects as silos collapses, 2- When educating new professionals, providing them with the resilience and resistance necessaries to deal with a problem is always mandatory, but now seems more relevant and 3.- collaborative work and contact with real practitioners in such an uncertain and changing environment is a challenge, but it is worth when considering the learning result and its potential.

Keywords: problem-based learning, hospitality recovery, collaborative learning, resilience

Procedia PDF Downloads 172

Authors: Danah Al-Selahi, Sulaiman Al-Ghunaim, Bashayer Sadiq, Fatma Al-Otaibi, Ali Ameen

Abstract:

In light of recent changes affecting the Oil & Gas Industry, optimization measures have become imperative for all companies globally, including Kuwait Oil Company (KOC). To keep abreast of the dynamic market, a detailed Integrated Asset Plan (IAP) was developed to drive optimization across the organization, which was facilitated through the in-house developed software “Well Inventory Data Entry” (WIDE). This comprehensive and integrated approach enabled centralization of all planned asset components for better well planning, enhancement of performance, and to facilitate continuous improvement through performance tracking and midterm forecasting. Traditionally, this was hard to achieve as, in the past, various legacy methods were used. This paper briefly describes the methods successfully adopted to meet the company’s objective. IAPs were initially designed using computerized spreadsheets. However, as data captured became more complex and the number of stakeholders requiring and updating this information grew, the need to automate the conventional spreadsheets became apparent. WIDE, existing in other aspects of the company (namely, the Workover Optimization project), was utilized to meet the dynamic requirements of the IAP cycle. With the growth of extensive features to enhance the planning process, the tool evolved into a centralized data-hub for all asset-groups and technical support functions to analyze and infer from, leading WIDE to become the reference two-year operational plan for the entire company. To achieve WIDE’s goal of operational efficiency, asset-groups continuously add their parameters in a series of predefined workflows that enable the creation of a structured process which allows risk factors to be flagged and helps mitigation of the same. This tool dictates assigned responsibilities for all stakeholders in a method that enables continuous updates for daily performance measures and operational use. The reliable availability of WIDE, combined with its user-friendliness and easy accessibility, created a platform of cross-functionality amongst all asset-groups and technical support groups to update contents of their respective planning parameters. The home-grown entity was implemented across the entire company and tailored to feed in internal processes of several stakeholders across the company. Furthermore, the implementation of change management and root cause analysis techniques captured the dysfunctionality of previous plans, which in turn resulted in the improvement of already existing mechanisms of planning within the IAP. The detailed elucidation of the 2 year plan flagged any upcoming risks and shortfalls foreseen in the plan. All results were translated into a series of developments that propelled the tool’s capabilities beyond planning and into operations (such as Asset Production Forecasts, setting KPIs, and estimating operational needs). This process exemplifies the ability and reach of applying advanced development techniques to seamlessly integrated the planning parameters of various assets and technical support groups. These techniques enables the enhancement of integrating planning data workflows that ultimately lay the founding plans towards an epoch of accuracy and reliability. As such, benchmarks of establishing a set of standard goals are created to ensure the constant improvement of the efficiency of the entire planning and operational structure.

Keywords: automation, integration, value, communication

Procedia PDF Downloads 124

Authors: Philipp Sommer, Amgad Agoub

Abstract:

The construction sector significantly impacts global CO₂ emissions, particularly through the energy usage of residential buildings. To address this, various governments, including Germany's, are focusing on reducing emissions via sustainable refurbishment initiatives. This study examines the application of machine learning (ML) to estimate energy demands dynamically in residential buildings and enhance the potential for large-scale sustainable refurbishment. A major challenge in Germany is the lack of extensive publicly labeled datasets for energy performance, as energy performance certificates, which provide critical data on building-specific energy requirements and consumption, are not available for all buildings or require on-site inspections. Conversely, England and other countries in the European Union (EU) have rich public datasets, providing a viable alternative for analysis. This research adapts insights from these English datasets to the German context by developing a comprehensive data schema and calibration dataset capable of predicting building energy demand effectively. The study proposes a minimal feature set, determined through feature importance analysis, to optimize the ML model. Findings indicate that ML significantly improves the scalability and accuracy of energy demand forecasts, supporting more effective emissions reduction strategies in the construction industry. Integrating energy performance certificates into municipal heat planning in Germany highlights the transformative impact of data-driven approaches on environmental sustainability. The goal is to identify and utilize key features from open data sources that significantly influence energy demand, creating an efficient forecasting model. Using Extreme Gradient Boosting (XGB) and data from energy performance certificates, effective features such as building type, year of construction, living space, insulation level, and building materials were incorporated. These were supplemented by data derived from descriptions of roofs, walls, windows, and floors, integrated into three datasets. The emphasis was on features accessible via remote sensing, which, along with other correlated characteristics, greatly improved the model's accuracy. The model was further validated using SHapley Additive exPlanations (SHAP) values and aggregated feature importance, which quantified the effects of individual features on the predictions. The refined model using remote sensing data showed a coefficient of determination (R²) of 0.64 and a mean absolute error (MAE) of 4.12, indicating predictions based on efficiency class 1-100 (G-A) may deviate by 4.12 points. This R² increased to 0.84 with the inclusion of more samples, with wall type emerging as the most predictive feature. After optimizing and incorporating related features like estimated primary energy consumption, the R² score for the training and test set reached 0.94, demonstrating good generalization. The study concludes that ML models significantly improve prediction accuracy over traditional methods, illustrating the potential of ML in enhancing energy efficiency analysis and planning. This supports better decision-making for energy optimization and highlights the benefits of developing and refining data schemas using open data to bolster sustainability in the building sector. The study underscores the importance of supporting open data initiatives to collect similar features and support the creation of comparable models in Germany, enhancing the outlook for environmental sustainability.

Keywords: machine learning, remote sensing, residential building, energy performance certificates, data-driven, heat planning

Procedia PDF Downloads 25

Authors: Charaia Vakhtang

Abstract:

Global oil prices are seeing new bottoms every day. The prices have already collapsed beneath the psychological verge of 30 USD. This tendency would be fully acceptable for the Georgian consumers, but there is one detail: two our neighboring countries (one friendly and one hostile) largely depend on resources of these hydrocarbons. Namely, the ratio of Azerbaijan in Georgia’s total FDI inflows in 2014 marked 20%. The ratio reached 40% in the January to September 2015. Azerbaijan is Georgia’s leading exports market. Namely, in 2014 Georgia’s exports to Azerbaijan constituted 544 million USD, i.e. 19% in Georgia’s total experts. In the January to November period of 2015, the ratio exceeded 11%. Moreover, Azerbaijan is Georgia’s strategic partner country as part of many regional projects that are designated for long-term perspectives. For example, the Baku-Tbilisi-Karsi railroad, the Black Sea terminal, preferential gas tariffs for Georgia and so on. The Russian economic contribution to the Georgian economy is also considerable, despite the losses the Russian hostile policy has inflicted to our country. Namely, Georgian emigrants are mainly employed in the Russian Federation and this category of Georgian citizens transfers considerable funds to Georgia every year. These transfers account for about 1 billion USD and consequently, these funds previously equalized to total FDI inflows. Moreover, despite the difficulties in the Russian market, Russia still remains a leader in terms of money transfers to Georgia. According to the last reports, money transfers from Russia to Georgia slipped by 276 million USD in 2015 compared to 2014 (-39%). At the same time, the total money transfers to Georgia in 2015 marked 1.08 billion USD, down 25% from 1.44 billion USD in 2014. This signifies the contraction in money transfers is by ¾ dependent on the Russian factor (in this case, contraction in oil prices and the Russian Ruble devaluation directly make negative impact on money transfers to Georgia). As to other countries, it is interesting that money transfers have also slipped from Italy (to 109 million USD from 121 million USD). Nevertheless, the country’s ratio in total money transfers to Georgia has increased to 10% from 8%. Money transfers to Georgia have increased by 22% (+18 million USD) from the USA. Money transfers have halved from Greece to 117 million USD from 205 million USD. As to Turkey, money transfers to Georgia from Turkey have increased by 1% to 69 million USD. Moreover, the problems with the national currencies of Russia and Azerbaijan, along with the above-mentioned developments, outline unfavorable perspectives for the Georgian economy. The depreciation of the national currencies of Azerbaijan and Russia is expected to bring unfavorable results for the Georgian economy. Even more so, the statement released by the Russian Finance Ministry on expected default is in direct relation to the welfare of the whole region and these tendencies will make direct and indirect negative impacts on Georgia’s economic indicators. Amid the economic slowdown in Armenia, Turkey and Ukraine, Georgia should try to enhance economic ties with comparatively stronger and flexible economies such as EU and USA. In other case, the Georgian economy will enter serious turbulent zone. We should make maximum benefit from the EU association agreement. It should be noted that the Russian economy slowdown that causes both regretful and happy moods in Georgia, will make negative impact on the Georgian economy. The same forecasts are made in relation to Azerbaijan. However, Georgia has many partner countries. Enhancement and development of the economic relations with these countries may maximally alleviate negative impacts from the declining economies. First of all, the EU association agreement should be mentioned as a main source for Georgia’s economic stabilization. It is the Georgian government‘s responsibility to successfully fulfill the EU association agreement requirements. In any case the imports must be replaced by domestic products and the exports should be stimulated through government support programs. The Authorities should ensure drawing more foreign investments and money resources, accumulating more tourism revenues and reducing external debts, budget expenditures should be balanced and the National Bank should carry out strict monetary policy. Moreover, the Government should develop a long-term state economic policy and carry out this policy at various Ministries. It is also of crucial importance to carry out constitutive policy and promote perspective directions on the domestic level.

Keywords: oil prices, economic growth, foreign direct investments, international trade

Procedia PDF Downloads 251

Authors: Ali Kazemi

Abstract:

Those within the volatile and interconnected international economic markets, appropriately predicting market trends, hold substantial fees for traders and financial establishments. Traditional device mastering strategies have made full-size strides in forecasting marketplace movements; however, monetary data's complicated and networked nature calls for extra sophisticated processes. This observation offers a groundbreaking method for monetary marketplace prediction that leverages the synergistic capability of Graph Convolutional Networks (GCNs) and Long Short-Term Memory (LSTM) networks. Our suggested algorithm is meticulously designed to forecast the traits of inventory market indices and cryptocurrency costs, utilizing a comprehensive dataset spanning from January 1, 2015, to December 31, 2023. This era, marked by sizable volatility and transformation in financial markets, affords a solid basis for schooling and checking out our predictive version. Our algorithm integrates diverse facts to construct a dynamic economic graph that correctly reflects market intricacies. We meticulously collect opening, closing, and high and low costs daily for key inventory marketplace indices (e.g., S&P 500, NASDAQ) and widespread cryptocurrencies (e.g., Bitcoin, Ethereum), ensuring a holistic view of marketplace traits. Daily trading volumes are also incorporated to seize marketplace pastime and liquidity, providing critical insights into the market's shopping for and selling dynamics. Furthermore, recognizing the profound influence of the monetary surroundings on financial markets, we integrate critical macroeconomic signs with hobby fees, inflation rates, GDP increase, and unemployment costs into our model. Our GCN algorithm is adept at learning the relational patterns amongst specific financial devices represented as nodes in a comprehensive market graph. Edges in this graph encapsulate the relationships based totally on co-movement styles and sentiment correlations, enabling our version to grasp the complicated community of influences governing marketplace moves. Complementing this, our LSTM algorithm is trained on sequences of the spatial-temporal illustration discovered through the GCN, enriched with historic fee and extent records. This lets the LSTM seize and expect temporal marketplace developments accurately. Inside the complete assessment of our GCN-LSTM algorithm across the inventory marketplace and cryptocurrency datasets, the version confirmed advanced predictive accuracy and profitability compared to conventional and opportunity machine learning to know benchmarks. Specifically, the model performed a Mean Absolute Error (MAE) of 0.85%, indicating high precision in predicting day-by-day charge movements. The RMSE was recorded at 1.2%, underscoring the model's effectiveness in minimizing tremendous prediction mistakes, which is vital in volatile markets. Furthermore, when assessing the model's predictive performance on directional market movements, it achieved an accuracy rate of 78%, significantly outperforming the benchmark models, averaging an accuracy of 65%. This high degree of accuracy is instrumental for techniques that predict the course of price moves. This study showcases the efficacy of mixing graph-based totally and sequential deep learning knowledge in economic marketplace prediction and highlights the fee of a comprehensive, records-pushed evaluation framework. Our findings promise to revolutionize investment techniques and hazard management practices, offering investors and economic analysts a powerful device to navigate the complexities of cutting-edge economic markets.

Keywords: financial market prediction, graph convolutional networks (GCNs), long short-term memory (LSTM), cryptocurrency forecasting

Procedia PDF Downloads 30