Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Christian spirituality --- Christian pastoral theology --- Art

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

De auteurs hebben de turnoefeningen stap per stap opgebouwd, en ze bieden een brede variatie, zodat elke turner op zijn niveau kan oefenen. De materiaalopstellingen zijn efficiënt en elke oefening wordt aangevuld met didactische tips en helperstechnieken. Zowel klassieke turnoefeningen als nieuwe trends zoals freerunnen komen aan bod. Voor mij is dit boek een echte topper!" Vind jij turnen fantastisch en wil je in het turnen grenzen verleggen? Turn IT on geeft de perfecte oplossing. In dit boek werken de auteurs Bart Thijs en Robin De Vil via bewegingsfamilies een logische turnopbouw uit met focus op sprongen. Steunen, rotaties en slingeren komen in een ander boekdeel aan bod. De auteurs hechten veel belang aan een speelse, gevarieerde én veilige aanpak van de spel- en oefenvormen. Ze kijken over de muur van de turnzaal en gaan ook freerunnen. Vanuit hun jarenlange ervaring en tomeloze plezier in het turnen dompelen de auteurs je onder in het brede spectrum van recreatief turnen en freerunnen. Ze geven antwoorden op de meest uiteenlopende vragen die je je als lesgever, trainer of begeleider al eens gesteld hebt, zoals: hoe motiveer je je deelnemers? Hoe ga je om met niveauverschillen en grote groepen? Hoe zorg je voor een snelle succesbeleving, zowel individueel als in groep? Leg je graag de link naar andere sporten? Wil je je deelnemers zelfstandig maar toch gestructureerd laten werken? Turn IT on! Dat willen we iedereen aanraden, of je nu lesgeeft in het basis- of secundair onderwijs, of trainingen begeleidt in de turnclub of een recreatieve vereniging. Met Turn IT on. Turnen en freerunnen vanuit sprongen ga je off- én online aan de slag. De vele foto"s en zaalopstellingen nemen je onmiddellijk mee in de oefeningen. En de filmpjes, die via QR-codes scanbaar zijn met smartphone of tablet, helpen je de bewegingen duidelijk aan te leren en te laten uitvoeren. Dennis Goossens, Belgisch turner gespecialiseerd in de ringen, achtste op de Olympische Spelen van Rio (2016). "Ik vind turnen pure FUN. Het is grenzen verleggen én beleven. En dat is exact de focus van dit boek. Het zet je op weg naar turnlessen waarin zowel de doorsneeleerling als de gevorderde recreatieve gymnast uitgedaagd wordt.

Turnen --- Recreation. Games. Sports. Corp. expression --- gymnastiek --- springen --- Didactics of sports and games --- freerunning --- steunsprongen --- salto --- 616.2 --- sport --- bewegingsopvoeding --- turnen --- 616.8 --- Streksprong --- Hurksprong --- Spreidsprong --- Steunsprongen --- Rekstok --- Turnen - Atletiek --- PXL-Education 2018 --- didactiek van sport en spel --- 796.41 --- 796.41 Gymnastiek --- Gymnastiek --- Parkour --- Turnen ; didactiek --- Gymnastics --- Didactics of primary education --- Didactics of secundary education --- sporttraining --- lerarenopleiding --- sportclubs

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

In this internship we examine the possibility of clustering scientific abstracts on the basis of their content. A previous approach based these clusters on noun phrases (NPs) extracted from the ab-stract’s sentences, but it was found that they were rather inconsistent; some clusters were formed on the basis of their object of study while others occurred due to similar techniques that were em-ployed. Hence we devise an approach to automatically detect an abstract’s structure in order to be able to determine whether a NP occurs in the ‘introduction’, ‘methodology’ or ‘conclusion’ section. A Random Forest classifier with a wide range of structural, morpho-syntactic and lexical features performed best for this task. The accuracy of the model was fairly satisfactory (about 87%, although the recall for methodology sentences was quite low). However, the clusters that were formed on the basis of this classification were rather unsatisfactory. We suggest that the most likely explana-tion is the lack of any criterion to distinguish ‘irrelevant’ extracted NPs from others, and propose several solutions to this problem.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

In this thesis, I explore whether graph representational learning can enhance the detection of money laundering. Money laundering can be represented as a pattern in a graph where transactions are mapped as edges and the receiving or sending entities as nodes. These patterns can be captured using graph representational learning and be used in a classifier to predict if someone is involved in money laundering. I've explored various methods for graph representational learning and assessed them based on requirements derived from this use case. The GraphSAGE algorithm which is based on a graph convolutional neural network, is selected as most promising. This algorithm builds a representation of a neighbourhood per hop. At every hop, the local neighbourhood of a node is aggregated into one vector, reduced in size with a convolutional layer and is used as input for the next hop. The reduced vector of a neighbourhood in the final hop is called the node embedding and contains condensed information of the nodes local graph neighbourhood. I've created node embeddings for the customers using three different approaches; the GraphSAGE algorithm, a variant of the GraphSAGE which includes edge labels and direction in combination with a deterministic sampling approach and a graph auto encoder using the sampling strategy of GraphSAGE. I've used these embeddings together with the classical feature set, derived using domain knowledge, and have trained a random forest classifier to assess if the embeddings can enhance the detection performance. The results show no significant improvement in performance between the classifier trained including the embedding features and excluding. This is likely due to the limited number of positive labels available for training. To strengthen this assumption, I've trained a classifier to predict whether the node is a retail customer or a corporate customer using only the embeddings. This problem is better balanced and contains more labels. The performance of this classifier has a precision recall auc of >99%. Secondly, I've clustered the embedding of all corporate customers and compared the cluster assignment with the industry type, assuming that corporates with the same industry type have similar transaction patterns and therefore a similar local graph neighbourhood. The clustering shows a significant correlation with the industry type. From these experiments, I conclude that embeddings can be effective in condensing the information of a customers' local graph neighbourhood.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Contemporary science mapping is usually done by clustering or is based on citations. In addition, the journals in which they are published, could serve as an indication of the scientific field the publication belongs to as well. In this paper, a multi-label classification model based on the content of the publications is developed. This means a predefined framework is at hand, unlike the clustering methods, and the content of the article itself is used instead of citations or journals. This paper is the progress report of an internship at ECOOM, Leuven. To get started, some traditional methods, such as Naive Bayes and logistic regression, are used to set up a baseline classifier. The final model is an LSTM neural network.