Pub Date : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281234
Julio Barzola-Monteses, Mayken Espinoza-Andaluz, Mónica Mite-León, Manuel Flores-Morán
Buildings have a dominant presence in energy consumption for the transition to clean energy. During 2017, construction and operation of buildings worldwide represented more than a third (36%) of final energy used and 40% of the emissions of carbon dioxide. Hence, there is great interest in reducing energy use in this sector, and energy efficiency in buildings to enhance energy performances is a suitable way. In this paper, black-box approaches based on artificial neural networks to predict the electric load of a selected educational building are proposed. The potential and robustness of long short-term memory (LSTM) applied to a dataset with a limited number of days of observations are analyzed. The results in our scenario showed that the LSTM surpasses in accuracy to other techniques such as feed-forward neural networks.
{"title":"Energy Consumption of a Building by using Long Short-Term Memory Network: A Forecasting Study","authors":"Julio Barzola-Monteses, Mayken Espinoza-Andaluz, Mónica Mite-León, Manuel Flores-Morán","doi":"10.1109/SCCC51225.2020.9281234","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281234","url":null,"abstract":"Buildings have a dominant presence in energy consumption for the transition to clean energy. During 2017, construction and operation of buildings worldwide represented more than a third (36%) of final energy used and 40% of the emissions of carbon dioxide. Hence, there is great interest in reducing energy use in this sector, and energy efficiency in buildings to enhance energy performances is a suitable way. In this paper, black-box approaches based on artificial neural networks to predict the electric load of a selected educational building are proposed. The potential and robustness of long short-term memory (LSTM) applied to a dataset with a limited number of days of observations are analyzed. The results in our scenario showed that the LSTM surpasses in accuracy to other techniques such as feed-forward neural networks.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116688133","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 : 2020-11-16DOI: 10.1109/sccc51225.2020.9281176
{"title":"SCCC 2020 Committees","authors":"","doi":"10.1109/sccc51225.2020.9281176","DOIUrl":"https://doi.org/10.1109/sccc51225.2020.9281176","url":null,"abstract":"","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"34 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123247681","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281276
R. León, Carola A. Blazquez, Vincent Depassier
The map-matching problem commonly arises when integrating position and other information from Global Navigation Satellite Systems (GNSS) such as GPS into a digital road map. This study presents improvements to an existing post-processing topological map-matching algorithm (TMMA) that successfully solves this problem. Both existing and improved TMMA were tested and compared regarding solution quality and computation time using GPS data collected from nine winter maintenance vehicle routes in Portage County, Wisconsin in the United States. On average, the results indicate an increase of 0.6% in the correct assignment of GPS points to the road network, and a decrease of 1% in the false negative (FN) cases (unmatched GPS points) when comparing the improved TMMA to the existing TMMA. Additionally, the improved TMMA can solve on average 1.3% more cases than the existing TMMA by assigning incorrect and FN points to correct road segments. Although enhanced results in terms of solution quality were obtained with the improved TMMA, the computation time is increased with this version of the TMMA due to additional steps that are incorporated in the resolution of the map-matching problem. Finally, the paired-sample T tests were conducted to identify statistical differences between both versions of the TMMA.
{"title":"Improvements of a Topological Map-Matching Algorithm in Post-Processing Mode","authors":"R. León, Carola A. Blazquez, Vincent Depassier","doi":"10.1109/SCCC51225.2020.9281276","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281276","url":null,"abstract":"The map-matching problem commonly arises when integrating position and other information from Global Navigation Satellite Systems (GNSS) such as GPS into a digital road map. This study presents improvements to an existing post-processing topological map-matching algorithm (TMMA) that successfully solves this problem. Both existing and improved TMMA were tested and compared regarding solution quality and computation time using GPS data collected from nine winter maintenance vehicle routes in Portage County, Wisconsin in the United States. On average, the results indicate an increase of 0.6% in the correct assignment of GPS points to the road network, and a decrease of 1% in the false negative (FN) cases (unmatched GPS points) when comparing the improved TMMA to the existing TMMA. Additionally, the improved TMMA can solve on average 1.3% more cases than the existing TMMA by assigning incorrect and FN points to correct road segments. Although enhanced results in terms of solution quality were obtained with the improved TMMA, the computation time is increased with this version of the TMMA due to additional steps that are incorporated in the resolution of the map-matching problem. Finally, the paired-sample T tests were conducted to identify statistical differences between both versions of the TMMA.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122210047","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281222
José Ignacio Abreu Salas, P. Mirabal
Perturbation-based heuristics for the median string problem performs successive refinements to an incumbent solution. Perturbations can be applied (i) one by one or (ii) multiple at a time. Algorithms in (i) seem to converge to noticeable best quality solutions but are much slower than those in (ii).In this paper, we proposed an algorithm performing several edit operations at a time with a better trade-off between quality and execution time. A profuse experimental evaluation suggests our approach can converge to much better solutions than former algorithms in (ii), even comparable with (i) in some cases. In counterpart, it is much faster than those in (i) but slower than those in (ii).
{"title":"Multi-edition approach for Median String Problem","authors":"José Ignacio Abreu Salas, P. Mirabal","doi":"10.1109/SCCC51225.2020.9281222","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281222","url":null,"abstract":"Perturbation-based heuristics for the median string problem performs successive refinements to an incumbent solution. Perturbations can be applied (i) one by one or (ii) multiple at a time. Algorithms in (i) seem to converge to noticeable best quality solutions but are much slower than those in (ii).In this paper, we proposed an algorithm performing several edit operations at a time with a better trade-off between quality and execution time. A profuse experimental evaluation suggests our approach can converge to much better solutions than former algorithms in (ii), even comparable with (i) in some cases. In counterpart, it is much faster than those in (i) but slower than those in (ii).","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128225851","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281213
Francisco Tobar M, Claudio E. Torres
Generative Adversarial Networks (GANs) for image generation of human faces have provided excellent results in recent years. However, we were able to identify a common problem among them: high memory usage in their training phase due to the convolutional encoder architecture used in these models. We address this issue by replacing the traditional convolutional layers in a model by what we call a Multi-Octave Convolution (M-OctConv) without modifying its architecture. An advantage of this method is that it can be easily combined with traditional memory reduction techniques, such as pruning. We evaluate our proposition on StarGAN model achieving up to 40% of memory usage reduction without affecting the quality of the generated images.
{"title":"Multi-OctConv: Reducing Memory Requirements in Image Generative Adversarial Networks","authors":"Francisco Tobar M, Claudio E. Torres","doi":"10.1109/SCCC51225.2020.9281213","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281213","url":null,"abstract":"Generative Adversarial Networks (GANs) for image generation of human faces have provided excellent results in recent years. However, we were able to identify a common problem among them: high memory usage in their training phase due to the convolutional encoder architecture used in these models. We address this issue by replacing the traditional convolutional layers in a model by what we call a Multi-Octave Convolution (M-OctConv) without modifying its architecture. An advantage of this method is that it can be easily combined with traditional memory reduction techniques, such as pruning. We evaluate our proposition on StarGAN model achieving up to 40% of memory usage reduction without affecting the quality of the generated images.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128367206","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281174
N. Esquivel, O. Nicolis, Billy Peralta Márquez
Vehicle theft represents one of the most frequent crimes in Chile and in the world. In this work, we propose an application of the GCLSTM (Graph-Convolutional Long Short Term Memory) neural network that combines a graph convolutional model with a LSTM for the prediction of vehicle thefts in the metropolitan region of Chile. The graph architecture considers the characteristics found in the neighbors to an area, assuming that the thefts of vehicles in nearby municipalities have similar patterns. For implementing the GCLSTM, first a smoothing technique based on LOESS regression was used for denoising the number of theft events for day, then the smoothed series of the last 30 days was considered as the input of the GCLSTM neural network for predicting the number of thefts in the following day. The results provided a better performance of the GCLSTM compared to a traditional LSTM, achieving an R2 of 0.86.
{"title":"Predicting Motor Vehicle Theft in Santiago de Chile using Graph-Convolutional LSTM","authors":"N. Esquivel, O. Nicolis, Billy Peralta Márquez","doi":"10.1109/SCCC51225.2020.9281174","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281174","url":null,"abstract":"Vehicle theft represents one of the most frequent crimes in Chile and in the world. In this work, we propose an application of the GCLSTM (Graph-Convolutional Long Short Term Memory) neural network that combines a graph convolutional model with a LSTM for the prediction of vehicle thefts in the metropolitan region of Chile. The graph architecture considers the characteristics found in the neighbors to an area, assuming that the thefts of vehicles in nearby municipalities have similar patterns. For implementing the GCLSTM, first a smoothing technique based on LOESS regression was used for denoising the number of theft events for day, then the smoothed series of the last 30 days was considered as the input of the GCLSTM neural network for predicting the number of thefts in the following day. The results provided a better performance of the GCLSTM compared to a traditional LSTM, achieving an R2 of 0.86.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127881912","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281282
A. González, José Lillo, Mario Inostroza-Ponta, Manuel Villalobos-Cid
The healthcare services must provide quality health safeguarding the efficient use of the resources. To evaluate technical efficiency performing fairly comparisons it is necessary to group the hospitals according to the type of patient treated: case-mix. Generally, this evaluation is performed by using the Related Groups for Diagnosis (DRG) system. Since only a few hospitals have implemented this system in Chile, the analysis of technical efficiency results limited. The Ministry of Health of Chile (MINSAL) has proposed an administrative categorisation for the public hospitals: high, medium and low complexity. However, it has not been studied if this definition is associated to the case-mix and if it can be used to study technical efficiency. In this work, we propose an ad-hoc genetic algorithm which combines filter feature selection and clustering strategies to determine if there is a set of features related to the case-mix that allow to reach the same categorisation proposed by the MINSAL. The results show that, although a small set of features is able to reach this categorisation by year, there is not enough evidence to establish a relationship with the case-mix. It is recommended that future technical efficiency analyses use new categorisations based on case-mix instead of the MINSAL categorisation.
{"title":"Evaluating the categorisation of the public hospitals in Chile according to case-mix complexity: a genetic algorithm approach","authors":"A. González, José Lillo, Mario Inostroza-Ponta, Manuel Villalobos-Cid","doi":"10.1109/SCCC51225.2020.9281282","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281282","url":null,"abstract":"The healthcare services must provide quality health safeguarding the efficient use of the resources. To evaluate technical efficiency performing fairly comparisons it is necessary to group the hospitals according to the type of patient treated: case-mix. Generally, this evaluation is performed by using the Related Groups for Diagnosis (DRG) system. Since only a few hospitals have implemented this system in Chile, the analysis of technical efficiency results limited. The Ministry of Health of Chile (MINSAL) has proposed an administrative categorisation for the public hospitals: high, medium and low complexity. However, it has not been studied if this definition is associated to the case-mix and if it can be used to study technical efficiency. In this work, we propose an ad-hoc genetic algorithm which combines filter feature selection and clustering strategies to determine if there is a set of features related to the case-mix that allow to reach the same categorisation proposed by the MINSAL. The results show that, although a small set of features is able to reach this categorisation by year, there is not enough evidence to establish a relationship with the case-mix. It is recommended that future technical efficiency analyses use new categorisations based on case-mix instead of the MINSAL categorisation.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132244783","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281247
Herwin Alayn Huillcen-Baca, Flor de Luz Palomino-Valdivia, J. Guizado, Y. P. Atencio, F. Tadeo
The change in ecosystems and the loss of biodiversity are global problems, one of the ecosystems with the most signifi-cant degradation is the high areas of the Andes, composed mostly of natural pastures. In the Andahuaylas province, Apurímac region, Peru, there is a high-impact Andean area for the collection of water for human consumption and irrigation; this area is called the Chumbao River Micro-basin. The problem is that this area is presenting essential changes in its surface, corresponding to natural pastures, especially of the species fescue (festuca dolycophylla) and paco (aciachne pulvinata). These changes do not have any estimates or studies that allow adequate decision-making in the adoption of preventive and corrective measures for the conservation of ecology, the environment, and water collection.Under this approach, this work proposes a method of estimating vegetation cover for those species, through the chromatic characteristics of each species, using high-resolution satellite images, extracted from the PERUSAT-1 satellite. The method consists of dividing the global satellite image into small images, converting them into the HSV color system (Hue, Saturation, and Value). Evaluating the range of chromatic characteristics of each species and performing range segmentation, subsequently fine-tuning the segmentation with morphological deformations and calculate the final area.The results obtained an accuracy of 91.56%, taking as a reference the result of the estimation of traditional vegetation cover; this result was tested in an area of 3824.45 m2 with the presence of both species. Therefore, our proposal is a reliable method for calculating vegetation cover and can be used for large surface areas, saving human and financial resources and with almost instantaneous results, compared to the traditional way.
{"title":"Vegetation cover estimation from high-resolution satellite images based on chromatic characteristics and image processing","authors":"Herwin Alayn Huillcen-Baca, Flor de Luz Palomino-Valdivia, J. Guizado, Y. P. Atencio, F. Tadeo","doi":"10.1109/SCCC51225.2020.9281247","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281247","url":null,"abstract":"The change in ecosystems and the loss of biodiversity are global problems, one of the ecosystems with the most signifi-cant degradation is the high areas of the Andes, composed mostly of natural pastures. In the Andahuaylas province, Apurímac region, Peru, there is a high-impact Andean area for the collection of water for human consumption and irrigation; this area is called the Chumbao River Micro-basin. The problem is that this area is presenting essential changes in its surface, corresponding to natural pastures, especially of the species fescue (festuca dolycophylla) and paco (aciachne pulvinata). These changes do not have any estimates or studies that allow adequate decision-making in the adoption of preventive and corrective measures for the conservation of ecology, the environment, and water collection.Under this approach, this work proposes a method of estimating vegetation cover for those species, through the chromatic characteristics of each species, using high-resolution satellite images, extracted from the PERUSAT-1 satellite. The method consists of dividing the global satellite image into small images, converting them into the HSV color system (Hue, Saturation, and Value). Evaluating the range of chromatic characteristics of each species and performing range segmentation, subsequently fine-tuning the segmentation with morphological deformations and calculate the final area.The results obtained an accuracy of 91.56%, taking as a reference the result of the estimation of traditional vegetation cover; this result was tested in an area of 3824.45 m2 with the presence of both species. Therefore, our proposal is a reliable method for calculating vegetation cover and can be used for large surface areas, saving human and financial resources and with almost instantaneous results, compared to the traditional way.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134175078","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281182
J. Munoz-Gama, Raul Alvarez-Esteban, José Montalva-Carmona, Jorge A. Baier
Low classroom engagement and distractions are important challenges of massive courses. The literature shows that those problems could decrease when the teacher addresses the students by their names. Learning all the names and faces for a small course is not a difficult task. However, in massive courses, this becomes a tedious and time-consuming task, making it not very practical. This work presents FAIS, a system for effectively learning students names and their corresponding faces, especially designed for massive courses. The system is based on memorization by repetition and gamification techniques, requiring less effort and including a playful perspective. A preliminary experience shows the learning of 105 students names and surnames, just during the daily commutes of the 10 days previous to the start of the course.
{"title":"FAIS: A System for Effectively Learning Students Names and Faces in Massive Courses","authors":"J. Munoz-Gama, Raul Alvarez-Esteban, José Montalva-Carmona, Jorge A. Baier","doi":"10.1109/SCCC51225.2020.9281182","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281182","url":null,"abstract":"Low classroom engagement and distractions are important challenges of massive courses. The literature shows that those problems could decrease when the teacher addresses the students by their names. Learning all the names and faces for a small course is not a difficult task. However, in massive courses, this becomes a tedious and time-consuming task, making it not very practical. This work presents FAIS, a system for effectively learning students names and their corresponding faces, especially designed for massive courses. The system is based on memorization by repetition and gamification techniques, requiring less effort and including a playful perspective. A preliminary experience shows the learning of 105 students names and surnames, just during the daily commutes of the 10 days previous to the start of the course.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"417 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133448907","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 : 2020-11-16DOI: 10.1109/SCCC51225.2020.9281196
Camilo Parra, C. Ponce, Rodrigo F. Salas
Reducing and preventing road traffic accidents is a major public health problem and a priority for many nations. In this paper, we seek to explore the performance of explainable machine learning models applied to the prediction of road traffic crashes using a dataset containing nearly three million records of this type of events and the conditions under which they occurred. To achieve this, the dataset US Accidents -A Countrywide Traffic Accident Dataset is used. First we will clean, standardize and reduce the data, then we will transform the time and location values using a geohashing library developed by Uber, later, we will increase our dataset to obtain events classified as ‘not an accident’ using web scraping techniques in the data sources of the original authors of the dataset. Then, we will evaluate the performance of different implementations of Random Forest and decision trees, we obtained a performance superior to 70% for the F1 score of these models. Finally, we conclude that weather conditions are strongly related to the car accident.
{"title":"Evaluating the Performance of Explainable Machine Learning Models in Traffic Accidents Prediction in California","authors":"Camilo Parra, C. Ponce, Rodrigo F. Salas","doi":"10.1109/SCCC51225.2020.9281196","DOIUrl":"https://doi.org/10.1109/SCCC51225.2020.9281196","url":null,"abstract":"Reducing and preventing road traffic accidents is a major public health problem and a priority for many nations. In this paper, we seek to explore the performance of explainable machine learning models applied to the prediction of road traffic crashes using a dataset containing nearly three million records of this type of events and the conditions under which they occurred. To achieve this, the dataset US Accidents -A Countrywide Traffic Accident Dataset is used. First we will clean, standardize and reduce the data, then we will transform the time and location values using a geohashing library developed by Uber, later, we will increase our dataset to obtain events classified as ‘not an accident’ using web scraping techniques in the data sources of the original authors of the dataset. Then, we will evaluate the performance of different implementations of Random Forest and decision trees, we obtained a performance superior to 70% for the F1 score of these models. Finally, we conclude that weather conditions are strongly related to the car accident.","PeriodicalId":117157,"journal":{"name":"2020 39th International Conference of the Chilean Computer Science Society (SCCC)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115619171","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：