Occupancy Forecasting Research Articles

Modelling and forecasting European Airbnb occupancy during the pandemic: The specific merits of panel-data and Markov-switching models

This study models European Airbnb occupancy from January 2017 to December 2020 for 43 European countries within an econometric panel-data framework. Besides typical economic covariates, the COVID-19 Stringency Index is employed to allow for the recent pandemic. The counterfactual observations created by two versions of the econometric model (one with and one without European aggregates of macroeconomic controls) are benchmarked against traditional univariate time-series models and a seasonal Markov-switching autoregression. For pandemic times, the latter performs very well in terms of several forecast accuracy measures due to its ability of (a) detecting the pandemic-induced structural break and (b) accounting for seasonal patterns in the data. The econometric models also predict well as they outperform the time-series models in the majority of cases. Hence, the panel-data approach is suitable when looking for economically interpretable Airbnb occupancy forecasts, even during the time of the pandemic.

SuPrNet: Super Proxy for 4D occupancy forecasting

Spacetime (4D) occupancy forecasting explicitly models the future development of self-driving vehicles (SDVs) themselves and their surrounding environment, which is highly beneficial for their downstream tasks. Currently, using point cloud forecasting as a proxy to self-supervisedly optimize the 4D occupancy forecasting framework has significant application prospects due to its independence from labeled data. However, existing methods have the following shortcomings: (1) Insufficient mining of spatiotemporal information in the scene. (2) Neglecting the extraction of regional deep features. (3) Inadequate learning ability for spatiotemporal predictive learning. Based on this, we propose SuPrNet, constructing a ”Super Proxy” to address the aforementioned issues. Specifically, SuPrNet first embeds the spatial and temporal information of the current frame into point-wise features through a 4D hybrid representation combined with multi-planar and grid features (4DPG). Then, a Hierarchical Encoder–Decoder Module (HED) with sparsity consideration is used to structure the point cloud and extract deep features. Thanks to the structured feature map representation, SuPrNet proceeds to use a spatiotemporal predictive learning module (STL) to implicitly predict future feature maps. Finally, we decode these feature maps to a specific size as the forecasting of future occupancy. The 4DPG, HED, and STL in the model are designed to address the aforementioned issues, and experiments on Nuscene and KITTI-Odometry show that our method achieves the best 4D occupancy forecasting and point cloud forecasting accuracy, surpassing previous state-of-the-art models by approximately 30%. Additionally, we have expanded and validated the application prospects of SuPrNet through various downstream tasks. Our project is available at https://github.com/AlanLiangC/4DPCF.

Decoding the future: Proposing an interpretable machine learning model for hotel occupancy forecasting using principal component analysis

Accurate hotel occupancy forecasting is vital for optimizing hotel revenue, yet interpretable machine learning tools lack extensive research. This paper presents a two-step approach utilizing historical and advanced booking data. Principal Components Analysis (PCA) groups similar patterns in booking curves, followed by a pickup forecasting model to predict occupancy. Evaluating the approach using real booking data from three European hotels (2018–2022), it outperformed two benchmarks: classical additive pickup and clustering-based pickup methods. Empirical results demonstrate the superiority of PCA-based methods across all hotels and forecasting horizons. Additionally, incorporating Average Daily Rates into PCA enhances daily hotel demand forecasts, offering potential for enhanced predictions with business operational information in a low-dimensional space.

Time Aware Long Short-Term Memory and Kronecker Gated Intelligent Transportation for Smart Car Parking

Technology desires to improve quality of life and impart citizen’s health as well as happiness. The concept of Internet of Things (IoT) refers to smart world where prevailing objects are said to be embedded and hence interact with each other (i.e., between objects and human beings) to achieve an objective. In the period of IoT as well as smart city, there is requirement for Intelligent Transport System-based (ITS) ingenious smart parking or car parking space prediction (CPSP) for more feasible cities. With the increase in population and mushroom growth in vehicles are bringing about several distinct economic as well as environmental issues. One of pivotal ones is optimal parking space identification. To address on this problem, in this work, Time-aware Long Short-Term Memory and Kronecker product Gated Recurrent Unit (TLSTM-KGRU) for smart parking in internet of transportation things is proposed. The TLSTM-KGRU method is split into two sections. In the first section, smart parking occupancy is derived using Time-aware Long Short-Term Memory (Time-aware LSTM) for Kuala Lumpur Convention Centre car parking sensor dataset. Following which the resultant smart car occupancy results are subjected to Linear Interpolations and Kronecker product Gated Recurrent Unit for predicting smart parking. When compared against other predictive methods such as SGRU-LSTM and CPSP using DELM, our experimental outcomes denote that TLSTM-KGRU method has improved performance for smart parking occupancy forecast as it not only enhances sensitivity and specificity but also reduces prediction time with minimum delay.

Enhanced prediction of parking occupancy through fusion of adaptive neuro-fuzzy inference system and deep learning models

While predicting parking occupancy is crucial for managing urban congestion, existing models often exhibit gaps in accuracy, uncertainty handling, and integration potential. This study introduces an innovative combination of adaptive neuro-fuzzy inference system (ANFIS) and deep learning (DL) techniques to address these shortcomings. Specifically, ANFIS is utilized for its proficiency in uncertainty representation via fuzzy set theory, whereas DL models excel in automatic feature learning, non-linear modeling, and identifying long-term dependencies in time-series parking data. By integrating ANFIS with recurrent neural networks (RNN), long short-term memory (LSTM), and gated recurrent units (GRU), we formulated the ANFIS-RNN, ANFIS-LSTM, and ANFIS-GRU fusion models, testing them on real-world parking datasets. Subsequent experiments highlighted the dominance of these fusion models over individual and benchmark counterparts. ANFIS-RNN achieved a 30.61% improvement in MSE, 16.70% in RMSE, 21.21% in MAE, 21.58% in MAPE, and a 1.03% elevation in R2 over the standalone RNN. The ANFIS-LSTM surpassed LSTM by 34.04% in MSE, 18.76% in RMSE, 26.16% in MAE, 27.71% in MAPE, with a 1.04% R2 increment. ANFIS-GRU exceeded GRU metrics by 27.54% in MSE, 14.85% in RMSE, 19.27% in MAE, 20.01% in MAPE, and boosted R2 by 1.03%. These outcomes underline the potential of integrated models in refining prediction precision. By leveraging the combined strengths of ANFIS and DL, this research offers a significant leap in parking occupancy forecasting. Its implications extend to data-centric urban planning and traffic regulation, marking a pivotal step for future endeavors in hybrid predictive modeling incorporating soft computing and deep learning paradigms.

Managing the classroom fund of an educational instituton

The article describes the experience of organizing work with classroom funds in educational institutions. We carried out a comparative analysis of foreign and national approaches to determining the criteria for the effective use of classroom funds as well as to modeling and automating processes related to classroom fund management.The research presents an author-proposed interpretation of setting classroom fund management objective. Moreover, it identifies a system of control actions, describes author-developed approaches to creating models, methods, and technologies for managing classroom funds. We have proposed several basic principles for adapting and implementing classroom fund management systems in educational institutions on the basis of the generalized experience of developing, implementing, and using classroom fund management technologies at various levels and with different models of education.Using the example of the Siberian State University of Geosystems and Technologies, we describe the experience of implementing a system that comprehensively solves the problems related to classroom fund management. They are as follows: keeping count, profiling, planning, occupancy forecasting, distribution, monitoring, and assessing the efficiency of using.The technologies and models of classroom fund management described in the article can be adapted and used to improve classroom fund management systems in organizations of various levels and models of education.

Improving the Prediction of Passenger Numbers in Public Transit Networks by Combining Short-Term Forecasts With Real-Time Occupancy Data

Passengers of public transportation nowadays expect reliable and accurate travel information. The need for occupancy information is becoming more prevalent in intelligent public transport systems as people started avoiding overcrowded vehicles during the COVID-19 pandemic. Furthermore, public transportation companies require accurate occupancy forecasts to improve service quality. We present a novel approach to improve the prediction of passenger numbers that enhances a day-ahead prediction with real-time data. We first train a baseline predictor on historical automatic passenger counting data. Next, we train a realtime model on the deviations between baseline prediction and observed values, thus capturing events not addressed by the baseline. For the forecast, we attempt to detect emerging patterns in real time and adjust the baseline prediction with deviations from the patterns. Our experiments with data from Germany show that the proposed model improves the forecast of the baseline model and is only outperformed by artificial neural networks in some instances. If the training sets only cover a limited period of up to four months, our approach outperforms competing methods. For larger training sets, there are mixed results in the sense that for some test cases, certain types of neural networks yield slightly better results, but our method still performs well with less training effort, is explainable along the whole prediction process and can be applied to existing prediction methods.

Judgmental Adjustments of Algorithmic Hotel Occupancy Forecasts: Does User Override Frequency Impact Accuracy at Different Time Horizons?

Judgmental adjustments of algorithmic predictions with the aim of improving demand forecast accuracy are a common revenue management practice. While empirical evidence on the impact of these user overrides is growing, little research attention has been given to the time horizon and the frequency in which these adjustments take place. Utilizing a multilevel regression model for repeated measures, 20,081,973 forecasts comprising seven different time horizons were analyzed. Data were collected from 1752 hotels of different hotel types belonging to 232 hotel chains in seven geographical regions. We find that the accuracy of algorithmic computer forecasts improves as time nears the date of stay and that the number of user overrides impacts this accuracy. The effect of the override frequency depends on the type of the forecasted demand and on the presence of special events. A higher number of user overrides is beneficial for group segment, but damaging for the transient segment. During special events periods, override frequency enhances accuracy.

Detecting occupancy behaviour of buildings through environment monitoring sensing

The goal of this research is to produce a realistic occupancy forecast as well as an ideal occupancy-based controller for increasing the performance of Heating Ventilation and Air-Conditioning [HVAC] systems. Reliable occupancy prediction is a critical enabler for demand-based HVAC control since it ensures that the HVAC system is not running unnecessarily while a room or zone is empty. We present simple yet successful occupancy prediction methods, as well as an approach for automatically setting temperature set-points, in this work. We present three alternative occupancy prediction algorithms based on historical occupancy measurements. The proposed approach uses historical occupancy profile and environmental room data to establish an occupancy identity strategy that can provide a stochastic model based on uncertain basis functions as an alternative. According to the findings, the proposed occupancy prediction algorithms could reach an accuracy of approximately 85% reliable prediction with few negative predictions.

Short-Term Forecasting of Off-Street Parking Occupancy

Information and communication technologies have opened the way to guide recent developments in the field of parking. In this paper these technologies are applied to model a decision support system that gives insight into 6-months ahead parking occupancy forecasts for 57 off-street parking locations in Amsterdam. An effect analysis was conducted into the influence of weather-, event-, parking tariff-, and public transport attributes on parking occupancy. The most influential factors on the parking occupancy were the scheduling of artistic and sports events, the addition of a public transport line, and the weather variables thunderstorm, average wind speed, temperature, precipitation, and sunshine duration. Parking tariffs did not significantly contribute to model performance, which could have been because of the lack of data and time variability in the parking tariffs of the examined parking locations. The forecasting algorithms compared were the seasonal naive model as a benchmark approach, the Box–Jenkins seasonal autoregressive integrated moving average with and without exogenous regressors (SARIMAX and SARIMA, respectively), exponential smoothing models, and the long short-term memory neural network. The SARIMAX model outperformed the other algorithms for the 6-months ahead forecasts according to the lowest root mean square error (RMSE). By including the event factor, the model improved by 24% based on the RMSE. Weather variables improved the predictive performance by 8%. Future studies could focus on the addition of more event variables, extension into an online model, and the impact of spatial–temporal features on parking occupancy.

Forecasting building occupancy: A temporal-sequential analysis and machine learning integrated approach

Building occupancy is the basis for building energy simulations, operations, and management. With the increasing need for energy conservation and the occupant-centric service of building energy systems, occupancy forecasting has become an essential input for simulations. These applications include model predictive control and demand response, with the potential to optimize the use of renewable energy sources. Based on recent research, occupancy forecasting tends to be based on occupancy data of the previous time step or continuous lagged dependent variables. Previous analysis demonstrated that building occupancy is innately temporal and sequential with seasonal features, which may be instructive for forecast research. This study proposes a temporal-sequential (TS) analysis and machine learning integrated approach for occupancy forecasting. Using hourly occupant data from 16 different buildings, we demonstrate that the proposed temporal-sequential analysis using a 1-week seasonal period with an artificial neural network structure (TS-week-ANN) outperforms other baseline methods.

Secondary Use of COVID-19 Symptom Incidence Among Hospital Employees as an Example of Syndromic Surveillance of Hospital Admissions Within 7 Days

Alternative methods for hospital occupancy forecasting, essential information in hospital crisis planning, are necessary in a novel pandemic when traditional data sources such as disease testing are limited. To determine whether mandatory daily employee symptom attestation data can be used as syndromic surveillance to estimate COVID-19 hospitalizations in the communities where employees live. This cohort study was conducted from April 2, 2020, to November 4, 2020, at a large academic hospital network of 10 hospitals accounting for a total of 2384 beds and 136 000 discharges in New England. The participants included 6841 employees who worked on-site at hospital 1 and lived in the 10 hospitals' service areas. Daily employee self-reported symptoms were collected using an automated text messaging system from a single hospital. Mean absolute error (MAE) and weighted mean absolute percentage error (MAPE) of 7-day forecasts of daily COVID-19 hospital census at each hospital. Among 6841 employees living within the 10 hospitals' service areas, 5120 (74.8%) were female individuals and 3884 (56.8%) were White individuals; the mean (SD) age was 40.8 (13.6) years, and the mean (SD) time of service was 8.8 (10.4) years. The study model had a MAE of 6.9 patients with COVID-19 and a weighted MAPE of 1.5% for hospitalizations for the entire hospital network. The individual hospitals had an MAE that ranged from 0.9 to 4.5 patients (weighted MAPE ranged from 2.1% to 16.1%). For context, the mean network all-cause occupancy was 1286 during this period, so an error of 6.9 is only 0.5% of the network mean occupancy. Operationally, this level of error was negligible to the incident command center. At hospital 1, a doubling of the number of employees reporting symptoms (which corresponded to 4 additional employees reporting symptoms at the mean for hospital 1) was associated with a 5% increase in COVID-19 hospitalizations at hospital 1 in 7 days (regression coefficient, 0.05; 95% CI, 0.02-0.07; P < .001). This cohort study found that a real-time employee health attestation tool used at a single hospital could be used to estimate subsequent hospitalizations in 7 days at hospitals throughout a larger hospital network in New England.

Real-World Implementation of an ICT-Based Platform to Promote Energy Efficiency

The energy efficiency requirements of most energy-consuming sectors have increased recently in response to climate change. For buildings, this means targeting both facility managers and building users with the aim of identifying potential energy savings and encouraging more energy-responsible behaviors. The Information and Communication Technology (ICT) platform developed in Horizon 2020 FEEdBACk project intends to fulfill these goals by enabling the optimization of energy consumption, generation, and storage and control of flexible devices without compromising comfort levels and indoor air quality parameters. This work aims to demonstrate the real-world implementation and functionality of the ICT platform composed of Load Disaggregation, Net Load Forecast, Occupancy Forecast, Automation Manager, and Behavior Predictor applications. Particularly, the results obtained by individual applications during the test phase are presented alongside the specific metrics used to evaluate their performance.

Rule-based scheduling of air conditioning using occupancy forecasting

Heating, ventilation and air conditioning systems represent considerable potential for energy savings, which can be realized through intelligent occupancy-centered control strategies. In this work, both supervised and unsupervised algorithms to forecast occupancy are proposed with the highest accuracies of 98.3% and 97.6%, respectively. Building on their output, a rule-based air conditioning scheduling technique is developed. As an example, a potential of 15.4% of energy savings is calculated using a dataset collected in a mid-size (4000 m2) building in Portugal.

Clustering and motif identification for occupancy-centric control of an air handling unit

Commercial and institutional buildings often operate at a fraction of their full occupancy, yet ventilation is frequently provided assuming maximum occupancy during working hours. Forecasting building occupancy can improve the operation of system level equipment by reducing chronic overventilation. However, developing actionable occupancy forecasting techniques for controls applications is subject to technological, security, and financial barriers. This paper proposes an occupancy forecasting technique that can be implemented in a controller given these constraints. Seven months of Wi-Fi, plug-in equipment and lighting electricity data were collected from an academic office building. Using Wi-Fi data, the building’s mean occupancy was demonstrated to be less than 25% of the 1000-person maximum. Representative daily occupancy profiles were produced by different clustering techniques. A classification tree was developed to determine motif occupancy profiles for day-ahead forecasting. Corresponding electrical profiles were taken to see if they followed the same trend as the occupancy profiles; 84.5% of days shared the same trends. The electrical data were fed through the classification tree, with a successful occupancy classification rate of 70.4% and error of 47 ± 69 persons at 95% confidence. A controls implementation for adaptive outdoor air damper actuation based on this technique is proposed for future work.

Ensuring skills are available in the right locations: are we there yet? A regional analysis of qualification gaps

ABSTRACT Matching skills supply and demand is critical for economic growth, competitiveness and inclusiveness. Yet, measuring skills is difficult. This paper develops an approach for measuring the gap between demand and supply of skills at the local level in the UK West Midlands region. It uses qualifications as a proxy for skills. By comparing the projected supply of qualifications with the forecast demand for qualifications from occupational forecasts, the analysis reveals significant skill mismatches with potential to constrain regional growth potential. Policy implications for connecting skills, transport and housing policy as part of local industrial strategies are explored.

Strategic zoning approach for urban areas: towards a shared transportation system

Abstract Investigating downstream freight demand is a prerequisite to accomplishing the overall strategic implementation of transportation systems. Machine learning has recently become widely applied in order to support decision-making in several logistic operational levels: travel/arrival time prediction, occupancy forecasting of logistic spaces, route optimization and so on. Nevertheless, strategic decision-making often overlooks flow tendencies forecasting. Targeting this perspective, the present paper aims at proposing an urban zoning approach based on time series forecasting of supply chain demand through clustering customers. To conduct our approach, we have selected a set of machine learning algorithms that are believed to be robust according to the literature and the achieved accuracy benchmarks. Considering real-life data-based computational results, a number of analytical insights are illustrated.

Recursive neural networks in hospital bed occupancy forecasting

BackgroundEfficient planning of hospital bed usage is a necessary condition to minimize the hospital costs. In the presented work we deal with the problem of occupancy forecasting in the scale of several months, with a focus on personnel’s holiday planning.MethodsWe construct a model based on a set of recursive neural networks, which performs an occupancy prediction using historical admission and release data combined with external factors such as public and school holidays. The model requires no personal information on patients or staff. It is optimized for a 60 days forecast during the summer season (May–September).ResultsAn average mean absolute percentage error (MAPE) of 6.24% was computed on 8 validation sets.ConclusionsThe proposed machine learning model has shown to be competitive to standard time-series forecasting models and can be recommended for incorporation in medium-size hospitals automatized scheduling and decision making.

Reservation Forecasting Models for Hospitality SMEs with a View to Enhance Their Economic Sustainability

In many tourism destinations, sustainability of the local economy leans on small and medium-sized hotels that are individually owned and operated by members of the community. Suffering from seasonality more than their big competitors, these hotels should undertake marketing initiatives to counteract wide demand fluctuations. Such initiatives are most effective if based on accurate occupancy forecasts, which must be performed at the individual hotel level. In this aim, the present paper suggests a demand forecasting approach adapted to specific features that characterize reservation data for small and medium-sized enterprises (SMEs) in the hospitality sector. The proposed framework integrates historical and advanced booking methods into a forecast combination with time-varying, performance-based weights. Whereas historical methods use only past observations about the number of guests recorded on a particular stay night to forecast future room occupancy (long-term perspective), advanced booking methods predict bookings-to-come based on partially accumulated data from reservations on hand (short-term perspective). In order to provide a possible solution to data sparsity issues that affect the application of advanced booking models to hospitality SMEs, a procedure that incorporates length-of-stay information directly into the reservation processing phase is also introduced. The methodology is tested on real time series of reservation data from three Italian hotels, located either in a city center (Milan) or in a typical destination for seasonal holidays (Lake Maggiore). Model parameters are calibrated on a training dataset and the accuracy of the occupancy forecasts is evaluated on a holdout sample. The results validate earlier findings about combinations of long-term and short-term forecasts and, in addition, show that using performance-based weights improves the quality of forecasts. Reducing the risk of large forecast failures, the proposed methodology can indeed have practical implications for the design and implementation of effective demand-side policies in hospitality SMEs. These policies are expected to provide a competitive advantage that can be crucial to the sustainability of small establishments in a context of growing global tourism.

Development of Fuzzy Time Series Model for Hotel Occupancy Forecasting

Receiving appropriate forecast accuracy is important in many countries’ economic activities, and developing effective and precise time series model is critical issue in tourism demand forecasting. In this paper, fuzzy rule-based system model for hotel occupancy forecasting is developed by analyzing 40 months’ time series data and applying fuzzy c-means clustering algorithm. Based on the values of root mean square error and mean absolute percentage error which are metrics for measuring forecast accuracy, it is defined that the model with 7 clusters and 4 inputs is the optimal forecasting model for hotel occupancy.