Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900767
Athanasios Christopoulos, M. Conrad, Mitul Shukla
The successful employment of Virtual Reality Environments in distance education contexts led to the development of various frameworks and taxonomies related to the Virtual-Learning approach. However, when it comes to Blended or ‘Hybrid’ Virtual Learning (HVL) scenarios, where the learners are concurrently co-present both in the physical and in the virtual environment, the lines are hard to be drawn as this has been a relatively unexplored area. Considering the aforementioned change in the setup of the educational context, different implications, challenges and outcomes are expected to be observed. Motivated by this shortcoming, we conducted a series of experiments with Computer Science and Technology students and investigated the impact of interactions on learners’ motivation to engage with the 3D virtual world and the educational activities by extension. In this paper, we discuss students’ preconceptions towards the inclusion of 3D Virtual Learning Environments in the context of their studies and further elicit their thoughts related to the impact of the ‘hybrid’ interactions. In addition, we investigate the educational value of different Non-Player Characters (Pedagogical Agents) and their impact on the attractiveness of the virtual world and the educational tasks. The concluding remarks provide guidance to educators and instructional designers who work in such setups or consider to employ Pedagogical Agents. To this end, employing Pedagogical Agents requires careful consideration as they need to be meaningful and fully incorporated in the learner’s task. Another take-away message concerns the elements that foster a situated learning experience as they are associated with immersive experiences.
{"title":"What Does the Pedagogical Agent Say?","authors":"Athanasios Christopoulos, M. Conrad, Mitul Shukla","doi":"10.1109/IISA.2019.8900767","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900767","url":null,"abstract":"The successful employment of Virtual Reality Environments in distance education contexts led to the development of various frameworks and taxonomies related to the Virtual-Learning approach. However, when it comes to Blended or ‘Hybrid’ Virtual Learning (HVL) scenarios, where the learners are concurrently co-present both in the physical and in the virtual environment, the lines are hard to be drawn as this has been a relatively unexplored area. Considering the aforementioned change in the setup of the educational context, different implications, challenges and outcomes are expected to be observed. Motivated by this shortcoming, we conducted a series of experiments with Computer Science and Technology students and investigated the impact of interactions on learners’ motivation to engage with the 3D virtual world and the educational activities by extension. In this paper, we discuss students’ preconceptions towards the inclusion of 3D Virtual Learning Environments in the context of their studies and further elicit their thoughts related to the impact of the ‘hybrid’ interactions. In addition, we investigate the educational value of different Non-Player Characters (Pedagogical Agents) and their impact on the attractiveness of the virtual world and the educational tasks. The concluding remarks provide guidance to educators and instructional designers who work in such setups or consider to employ Pedagogical Agents. To this end, employing Pedagogical Agents requires careful consideration as they need to be meaningful and fully incorporated in the learner’s task. Another take-away message concerns the elements that foster a situated learning experience as they are associated with immersive experiences.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123069964","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900693
Dimitris Mantzekis, M. Savelonas, S. Karkanis, E. Spyrou
Recurrent neural networks are an obvious choice for driving behavior analysis by means of time series of measurements, obtained either from telematics or mobile phone sensors. This work investigates such an application, employing two popular recurrent neural networks, i.e. long short-term memory networks and gated recurrent unit networks, as well as 1D convnets. Experiments are performed on a dataset comprising time series of measurements for four different types of driving. The results lead to the conclusion that gated recurrent unit networks achieve the highest classification accuracy, whereas they are more efficient than long short-term memory networks. Moreover, dropout and recurrent dropout lead to an approximately 3% increase with respect to classification accuracy. Naturally, 1D convnets are a more efficient neural network alternative at the cost of significantly lower classification accuracy.
{"title":"RNNs for Classification of Driving Behaviour","authors":"Dimitris Mantzekis, M. Savelonas, S. Karkanis, E. Spyrou","doi":"10.1109/IISA.2019.8900693","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900693","url":null,"abstract":"Recurrent neural networks are an obvious choice for driving behavior analysis by means of time series of measurements, obtained either from telematics or mobile phone sensors. This work investigates such an application, employing two popular recurrent neural networks, i.e. long short-term memory networks and gated recurrent unit networks, as well as 1D convnets. Experiments are performed on a dataset comprising time series of measurements for four different types of driving. The results lead to the conclusion that gated recurrent unit networks achieve the highest classification accuracy, whereas they are more efficient than long short-term memory networks. Moreover, dropout and recurrent dropout lead to an approximately 3% increase with respect to classification accuracy. Naturally, 1D convnets are a more efficient neural network alternative at the cost of significantly lower classification accuracy.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"32 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116126611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900725
A. Tsanousa, G. Meditskos, S. Vrochidis, Y. Kompatsiaris
Following the technological advancement and the constantly emerging assisted living applications, sensor-based activity recognition research receives great attention. Until recently, the majority of relevant research involved extracting knowledge out of single modalities, however, when individual sensors performances are not satisfactory, combining information from multiple sensors can be of use and improve the activity recognition rate. Early and late fusion classifier strategies are usually employed to successfully merge multiple sensors. This paper proposes a novel framework for combining accelerometers and gyroscopes at decision level, in order to recognize human activity. More specifically, we propose a weighted late fusion framework that utilizes the detection rate of a classifier. Furthermore, we propose the modification of an already existing class-based weighted late fusion framework. Experimental results on a publicly available and widely used dataset demonstrated that the combination of accelerometer and gyroscope under the proposed frameworks improves the classification performance.
{"title":"A Weighted Late Fusion Framework for Recognizing Human Activity from Wearable Sensors","authors":"A. Tsanousa, G. Meditskos, S. Vrochidis, Y. Kompatsiaris","doi":"10.1109/IISA.2019.8900725","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900725","url":null,"abstract":"Following the technological advancement and the constantly emerging assisted living applications, sensor-based activity recognition research receives great attention. Until recently, the majority of relevant research involved extracting knowledge out of single modalities, however, when individual sensors performances are not satisfactory, combining information from multiple sensors can be of use and improve the activity recognition rate. Early and late fusion classifier strategies are usually employed to successfully merge multiple sensors. This paper proposes a novel framework for combining accelerometers and gyroscopes at decision level, in order to recognize human activity. More specifically, we propose a weighted late fusion framework that utilizes the detection rate of a classifier. Furthermore, we propose the modification of an already existing class-based weighted late fusion framework. Experimental results on a publicly available and widely used dataset demonstrated that the combination of accelerometer and gyroscope under the proposed frameworks improves the classification performance.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124851355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900690
Dionisios N. Sotiropoulos, Ifigeneia Georgoula, Christos Bilanakos
This paper presents a non-linear optimization methodology for determining the Nash Equilibrium (NE) solutions of a non-cooperative two-player game. Each player, in particular, is trying to maximize a rational profit function within a continuous action space. The game arises in the context of a duopolistic network environment where two identical rival firms are competing to maximize their influence over a single consumer. Specifically, we consider a weighted and strongly connected network which mediates the opinion formation processes concerning the perceived qualities of their products. Obtaining the NE solutions for such a game is an extremely difficult task which cannot be analytically addressed, even if additional simplifying assumptions are imposed on the exogenous parameters of the model. Our approach, obtains the required NE solutions by combining the Karush-Kuhn-Tucker (KKT) conditions associated with the original optimization tasks into a single-objective nonlinear maximization problem under nonlinear constrains. The resulting optimization problem is, ultimately, solved through the utilization of the Sequential Quadratic Programming (SQP) algorithm which constitutes a state-of-the-art method for nonlinear optimization problems. The validity of our work is justified through the conduction of a series of experiments in which we simulated the best response-based dynamical behaviour of the two agents in the network that make strategic decisions. Juxtaposing the intersection points of the acquired best response curves against the NE solutions obtained by the proposed non-linear optimization methodology verifies that the corresponding solution points coincide.
{"title":"Optimal Duopolistic Competition Strategies in Social Networks","authors":"Dionisios N. Sotiropoulos, Ifigeneia Georgoula, Christos Bilanakos","doi":"10.1109/IISA.2019.8900690","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900690","url":null,"abstract":"This paper presents a non-linear optimization methodology for determining the Nash Equilibrium (NE) solutions of a non-cooperative two-player game. Each player, in particular, is trying to maximize a rational profit function within a continuous action space. The game arises in the context of a duopolistic network environment where two identical rival firms are competing to maximize their influence over a single consumer. Specifically, we consider a weighted and strongly connected network which mediates the opinion formation processes concerning the perceived qualities of their products. Obtaining the NE solutions for such a game is an extremely difficult task which cannot be analytically addressed, even if additional simplifying assumptions are imposed on the exogenous parameters of the model. Our approach, obtains the required NE solutions by combining the Karush-Kuhn-Tucker (KKT) conditions associated with the original optimization tasks into a single-objective nonlinear maximization problem under nonlinear constrains. The resulting optimization problem is, ultimately, solved through the utilization of the Sequential Quadratic Programming (SQP) algorithm which constitutes a state-of-the-art method for nonlinear optimization problems. The validity of our work is justified through the conduction of a series of experiments in which we simulated the best response-based dynamical behaviour of the two agents in the network that make strategic decisions. Juxtaposing the intersection points of the acquired best response curves against the NE solutions obtained by the proposed non-linear optimization methodology verifies that the corresponding solution points coincide.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127606995","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900749
C. Lytridis, E. Vrochidou, George K. Sidiropoulos, G. Papakostas, V. Kaburlasos, Efi Kourampa, Elpida Karageorgiou
Speech is one of the major human-machine interaction modalities and it is especially important in the case of special education using social robots. Although modern speech recognition engines can effectively deal with normal human-robot conversations, there are instances in special education where additional word detection and word comparison capabilities are needed to run in parallel with the typical conversation flow. This paper investigates the efficiency of a word detection method based on intervals’ numbers.
{"title":"Audio Signal Recognition Based on Intervals’ Numbers (INs) Classification Techniques","authors":"C. Lytridis, E. Vrochidou, George K. Sidiropoulos, G. Papakostas, V. Kaburlasos, Efi Kourampa, Elpida Karageorgiou","doi":"10.1109/IISA.2019.8900749","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900749","url":null,"abstract":"Speech is one of the major human-machine interaction modalities and it is especially important in the case of special education using social robots. Although modern speech recognition engines can effectively deal with normal human-robot conversations, there are instances in special education where additional word detection and word comparison capabilities are needed to run in parallel with the typical conversation flow. This paper investigates the efficiency of a word detection method based on intervals’ numbers.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"456 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115100303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900695
P. Schirmer, I. Mporas, M. Paraskevas
In this paper we evaluate several well-known and widely used machine learning algorithms for regression in the energy disaggregation task. Specifically, the Non-Intrusive Load Monitoring approach was considered and the K-Nearest-Neighbours, Support Vector Machines, Deep Neural Networks and Random Forest algorithms were evaluated across five datasets using seven different sets of statistical and electrical features. The experimental results demonstrated the importance of selecting both appropriate features and regression algorithms. The best performance in terms of energy disaggregation accuracy was achieved by the Random Forest regression algorithm.
{"title":"Evaluation of Regression Algorithms and Features on the Energy Disaggregation Task","authors":"P. Schirmer, I. Mporas, M. Paraskevas","doi":"10.1109/IISA.2019.8900695","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900695","url":null,"abstract":"In this paper we evaluate several well-known and widely used machine learning algorithms for regression in the energy disaggregation task. Specifically, the Non-Intrusive Load Monitoring approach was considered and the K-Nearest-Neighbours, Support Vector Machines, Deep Neural Networks and Random Forest algorithms were evaluated across five datasets using seven different sets of statistical and electrical features. The experimental results demonstrated the importance of selecting both appropriate features and regression algorithms. The best performance in terms of energy disaggregation accuracy was achieved by the Random Forest regression algorithm.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117083591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900736
Anna-Maria Velentza, Antonios Nikitakis, Konstantinos Alketas-Oungrinis, E. Economou
Virtual Reality Learning Environments (VLE) are a promising approach for the 21st century classroom. Therefore, it is important to identify the best way that these learning environments will enhance students’ cognitive functions. In this paper, we investigate the effect of lighting conditions, within a VLE, on several cognitive functions while we introduce the novel idea of transformable luminance conditions. We have been inspired by evidence from cognitive and environmental studies, from real environments, that lighting conditions affect people’s memory, attention and executive functions. Our transformable luminance approach, when applied in a virtual university amphitheatre, is benchmarked against the conventional approach of a single luminance VLE (either high or low). The first main outcome of our study is that traditional single luminance lighting conditions exert the same effect in both real and virtual learning environments. That is, high luminance enhances sustained attention and short term memory, while low luminance enhances long term memory. The results also indicate that our novel transformable luminance approach enhances student performance in executive tasks, compared to any single luminance condition (i.e. low or high).
{"title":"Transformable Lighting Conditions in Learning VR Environments","authors":"Anna-Maria Velentza, Antonios Nikitakis, Konstantinos Alketas-Oungrinis, E. Economou","doi":"10.1109/IISA.2019.8900736","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900736","url":null,"abstract":"Virtual Reality Learning Environments (VLE) are a promising approach for the 21st century classroom. Therefore, it is important to identify the best way that these learning environments will enhance students’ cognitive functions. In this paper, we investigate the effect of lighting conditions, within a VLE, on several cognitive functions while we introduce the novel idea of transformable luminance conditions. We have been inspired by evidence from cognitive and environmental studies, from real environments, that lighting conditions affect people’s memory, attention and executive functions. Our transformable luminance approach, when applied in a virtual university amphitheatre, is benchmarked against the conventional approach of a single luminance VLE (either high or low). The first main outcome of our study is that traditional single luminance lighting conditions exert the same effect in both real and virtual learning environments. That is, high luminance enhances sustained attention and short term memory, while low luminance enhances long term memory. The results also indicate that our novel transformable luminance approach enhances student performance in executive tasks, compared to any single luminance condition (i.e. low or high).","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116080059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900704
S. Panda, Chittotosh Ganguly, S. Chakrabarti
In an attempt to propose a closer model of a biological neuron, various artificial neural models have been reported in the literature. Very few reported articles are available which consider the time-varying synaptic weights of the model. Hence there is further scope to develop and investigate alternative improved spiking neural models which will better represent the activities of a biological neuron. With this motivation, the synaptic weight of the conventional integrate and fire (CIF) model is considered as gamma distributed time-varying nature. Further, for spike generation at the output of the model, different thresholds are employed. The gamma distribution in weight is assumed to take into account the temporal behavior of the synapse. To assess the performance of the proposed model, statistical properties such as similarity indices of the output sequence, mean and variance of normalized similarity indices (NSI) are obtained from simulation-based experiments and are compared.
{"title":"Spiking Neuron Model with Gamma-distributed Synaptic Weights for Different Thresholds","authors":"S. Panda, Chittotosh Ganguly, S. Chakrabarti","doi":"10.1109/IISA.2019.8900704","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900704","url":null,"abstract":"In an attempt to propose a closer model of a biological neuron, various artificial neural models have been reported in the literature. Very few reported articles are available which consider the time-varying synaptic weights of the model. Hence there is further scope to develop and investigate alternative improved spiking neural models which will better represent the activities of a biological neuron. With this motivation, the synaptic weight of the conventional integrate and fire (CIF) model is considered as gamma distributed time-varying nature. Further, for spike generation at the output of the model, different thresholds are employed. The gamma distribution in weight is assumed to take into account the temporal behavior of the synapse. To assess the performance of the proposed model, statistical properties such as similarity indices of the output sequence, mean and variance of normalized similarity indices (NSI) are obtained from simulation-based experiments and are compared.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"357 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122810425","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900748
Lambrini Seremeti, I. Kougias
Juris-informatics is a new research area which has emerged from the pervasive presence of sophisticated artifacts in the everyday scenery, for example every human activity, such as planning a journey, applying for a job, shopping, attending a course, etc. involves the use of intelligent ICT. Within these artificial intelligence environments, where the coexistence of living and non-living entities is their intrinsic characteristic, the main priority is ensuring a harmonious symbiosis of their stakeholders. This, in turn, leads to insoluble legal conflicts which are created by the fact that all the involved entities do not have the same liability against written law. The aim of this paper is to emphasize, by setting out some imaginative scenarios, yet posing real, everyday situations/questions, the lack of legal framework within artificial intelligence environments, which may lead to a dystopian society. Based on each such scenario, emerging legal issues are described that need to be tackled.
{"title":"Legal issues within ambient intelligence environments","authors":"Lambrini Seremeti, I. Kougias","doi":"10.1109/IISA.2019.8900748","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900748","url":null,"abstract":"Juris-informatics is a new research area which has emerged from the pervasive presence of sophisticated artifacts in the everyday scenery, for example every human activity, such as planning a journey, applying for a job, shopping, attending a course, etc. involves the use of intelligent ICT. Within these artificial intelligence environments, where the coexistence of living and non-living entities is their intrinsic characteristic, the main priority is ensuring a harmonious symbiosis of their stakeholders. This, in turn, leads to insoluble legal conflicts which are created by the fact that all the involved entities do not have the same liability against written law. The aim of this paper is to emphasize, by setting out some imaginative scenarios, yet posing real, everyday situations/questions, the lack of legal framework within artificial intelligence environments, which may lead to a dystopian society. Based on each such scenario, emerging legal issues are described that need to be tackled.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131133018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-07-01DOI: 10.1109/IISA.2019.8900680
A. Christidi, Constantinos Christidis, Tilemahos Manolatos, I. Hatzilygeroudis, E. Voyiatzaki
Business registries keep data from thousands of enterprises and offer a range of services that vary among different countries. Enterprises pay registration fees and fees per service. We propose to use machine learning to reveal existing payment behavior patterns and predict obligatory payments of enterprises to organizations responsible for business registries. To this end, we present a model of the problem suitable for applying machine learning methods.
{"title":"Modelling prediction of enterprises payment behavior for applying machine learning methods","authors":"A. Christidi, Constantinos Christidis, Tilemahos Manolatos, I. Hatzilygeroudis, E. Voyiatzaki","doi":"10.1109/IISA.2019.8900680","DOIUrl":"https://doi.org/10.1109/IISA.2019.8900680","url":null,"abstract":"Business registries keep data from thousands of enterprises and offer a range of services that vary among different countries. Enterprises pay registration fees and fees per service. We propose to use machine learning to reveal existing payment behavior patterns and predict obligatory payments of enterprises to organizations responsible for business registries. To this end, we present a model of the problem suitable for applying machine learning methods.","PeriodicalId":371385,"journal":{"name":"2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132213367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：