The paper explores the effectiveness of automated clustering in personalized applications based on data characteristics. It evaluates three clustering algorithms with various cluster numbers and subsets of characteristics. The study compares the accuracy of models in different clusters against original results and examines the algorithmic approaches and characteristic selections for optimal clustering performance. The research concludes that the proposed method aids in selecting appropriate clustering strategies and relevant characteristics for datasets. These insights may also guide further research on coaching approaches within applications.
DOCUMENT
Summary:A novel Smart Charging strategy, based on low base allowances per charger combined with 1. clustering of chargers on the same part of the grid and 2. dynamic non guaranteed allowance, is presented in this paper. This manner of Smart Charging will allow more than 3 times the amount of chargers to be installed in the existing grid, even when the grid is already congested. The system also improves the usage of available flexibility in EV charging compared to other Smart Charging strategies. The required algorithms are tested on public chargers in Amsterdam, in some of the most intensely used parts of the Dutch grid.
DOCUMENT
Machine learning models have proven to be reliable methods in classification tasks. However, little research has been done on classifying dwelling characteristics based on smart meter & weather data before. Gaining insights into dwelling characteristics can be helpful to create/improve the policies for creating new dwellings at NZEB standard. This paper compares the different machine learning algorithms and the methods used to correctly implement the models. These methods include the data pre-processing, model validation and evaluation. Smart meter data was provided by Groene Mient, which was used to train several machine learning algorithms. The models that were generated by the algorithms were compared on their performance. The results showed that Recurrent Neural Network (RNN) 2performed the best with 96% of accuracy. Cross Validation was used to validate the models, where 80% of the data was used for training purposes and 20% was used for testing purposes. Evaluation metrices were used to produce classification reports, which can indicate which of the models work the best for this specific problem. The models were programmed in Python.
DOCUMENT
The objective of this study was to generate groups of agri-food producers with high affinity in relation to their sustainable waste management practices. The aim of conforming these groups is the development of synergies, knowledge management, and policy- and decision-making by diverse stakeholders. A survey was conducted among the most experienced farmers in the region of Nuevo Urecho, Michoacán, Mexico, and a total of eight variables relating to sustainable waste management practices, agricultural food loss, and the waste generated at each stage of the production process were examined. The retrieved data were treated using the maximum inverse correspondence algorithm and the Galois Lattice was applied to generate clusters of highly affine producers. The results indicate 163 possible elements that generate the power set, and 31 maximum inverse correspondences were obtained. At this point, it is possible to determine the maximum number of relationships, called affinities. In general, all 15 considered farmers shared the measure of revaluation of food waste and 90% of the farmers shared affinity in measures related to ecological care and the proper management of waste. A practical implication of this study is the conformation of highly affine clusters for both policy and strategic decision-making.
LINK
The purpose of this study was to provide more insight into how the physical education (PE) context can be better tailored to the diverse motivational demands of secondary school students. Therefore, we examined how different constructs of student motivation in the context of PE combine into distinct motivational profiles, aiming to unveil motivational similarities and differences between students’ PE experiences. Participants were 2,562 Dutch secondary school students, aged 12–18, from 24 different schools. Students responded to questionnaires assessing their perception of psychological need satisfaction and frustration, and perceived mastery and performance climate in PE. In order to interpret the emerging profiles additional variables were assessed (i.e. demographic, motivational and PE-related variables). Two-step cluster analysis identified three meaningful profiles labelled as negative perceivers, moderate perceivers and positive perceivers. These three profiles differed significantly with regard to perceived psychological need satisfaction and frustration and their perception of the motivational climate. This study demonstrates that students can be grouped in distinct profiles based on their perceptions of the motivational PE environment. Consequently, the insights obtained could assist PE teachers in designing instructional strategies that target students’ differential motivational needs.
DOCUMENT
City authorities want to know how to match the charging infrastructures for electric vehicles with the demand. Using camera recognition algorithms from artificial intelligence we investigated the behavior of taxis at a charging stations and a taxi stand.
MULTIFILE
Nowadays, one of the major current health risks is excessive sitting during work hours. Furthermore, the coronavirus disease 2019 (COVID-19) pandemic and the corresponding government state of emergency forced many people to work from home. These constraints carried out an important change in the lifestyle of people; for instance, the proportion of sitting time in front of a computer during working hours has increased considerably worldwide, particularly through the implementation of teleworking.In order to motivate people to lead a less sedentary life, the Hanze University of Applied Sciences Groningen developed an automated recommender system. We investigated the possibility of automated coaching in order to increase physical activity and help people to reach their daily step goal. By monitoring people’s activity level and progress during the day, we predict personalized recommendations. The effect of these recommendations on the individual’s activity level forms the basis for a personalized coaching approach.Step count data is used to train a machine learning algorithm that estimates the hourly probability of the individual achieving the daily steps goal. The outcome of this prediction is combined with the effect of the type recommendation for the individual to deliver the best recommendation for the individual. To show the practical usefulness, we constructed a platform to manage the data, rules, machine learning algorithms and clustering of participants. Results of initial pilots using the platform and app have given insight in the performance of and challenges associated with algorithm selection and personal model generation for the coaching package caused by the nature of the data. Further research will therefore be done in optimizing machine learning algorithms and tuning for human datasets.
DOCUMENT
Machine learning models have proven to be reliable methods in classification tasks. However, little research has been conducted on the classification of dwelling characteristics based on smart meter and weather data before. Gaining insights into dwelling characteristics, which comprise of the type of heating system used, the number of inhabitants, and the number of solar panels installed, can be helpful in creating or improving the policies to create new dwellings at nearly zero-energy standard. This paper compares different supervised machine learning algorithms, namely Logistic Regression, Support Vector Machine, K-Nearest Neighbor, and Long-short term memory, and methods used to correctly implement these algorithms. These methods include data pre-processing, model validation, and evaluation. Smart meter data, which was used to train several machine learning algorithms, was provided by Groene Mient. The models that were generated by the algorithms were compared on their performance. The results showed that the Long-short term memory performed the best with 96% accuracy. Cross Validation was used to validate the models, where 80% of the data was used for training purposes and 20% was used for testing purposes. Evaluation metrics were used to produce classification reports, which indicates that the Long-short term memory outperforms the compared models on the evaluation metrics for this specific problem.
DOCUMENT
Despite the existence of various methods and abstraction techniques to reduce the privacy risk of process models generated by process mining algorithms, it is unclear how process mining stakeholders perceive privacy violations. In this pilot-study various process model visualisations were shown to 6 stakeholders of a travel expense claim process. While changing the abstraction levels of these visualisations, the stakeholders were asked whether they perceived a violation of their privacy. The results show that there are differences in how individual stakeholders perceive privacy violations of process models generated via process mining algorithms. Results differ per type of visualization, type of privacy risk reducing methods, changes of abstraction level and stakeholder role. To reduce the privacy risk, the interviewees suggested to include an authorization table in the process mining tool, communicate the goal of the analysis with all stakeholders, and validate the analysis with a privacy officer. It is suggested that future research focuses on discussing and validating process visualisations and privacy risk reducing methods and techniques with various process mining stakeholders in organisations. This is expected to reduce perceived violations and prevents developing techniques that are aimed at reducing privacy risk but are not considered as such by stakeholders.
MULTIFILE
Discussions on policy and management initiatives to facilitate individuals throughout working careers take place without sufficient insight into how career paths are changing, how these changes are related to a modernization of life course biographies, and whether this leads to increased labour market transitions. This paper asks how new, flexible labour market patterns can best be analyzed using an empirical, quantitative approach. The data used are from the career module of the Panel Study of Belgian Households (PSBH). This module, completed by almost 4500 respondents consists of retrospective questions tracing lengthy and even entire working life histories. To establish any changes in career patterns over such extended periods of time, we compare two evolving methodologies: Optimal Matching Analysis (OMA) and Latent Class Regression Analysis (LCA). The analyses demonstrate that both methods show promising potential in discerning working life typologies and analyzing sequence trajectories. However, particularities of the methods demonstrate that not all research questions are suitable for each method. The OMA methodology is appropriate when the analysis concentrates on the labour market statuses and is well equipped to make clear and interpretable differentiations if there is relative stability in career paths during the period of observation but not if careers become less stable. Latent Class has the strength of adopting covariates in the clustering allowing for more historically connected types than the other methodology. The clustering is denser and the technique allows for more detailed model fitting controls than OMA. However, when incorporating covariates in a typology, the possibilities of using the typology in later, causal, analyses is somewhat reduced.
MULTIFILE
