User generated content (UGC) provides abundant tourist information regarding destinations. The textual digital traces bring great opportunity along with great challenges. Text mining approaches including sentiment analysis, multiclass text classification, and network analysis are suitable for extracting the buried pattern under piles of unstructured data. We processed 18.721 reviews from worldwide tourists about Bali’s 15 topmost tourist attractions. This study uncovers the tourist perception through textual data using sentiment analysis to extract the positive and negative perceptions, and multiclass classification to extract the tourist cognitive concern for each destination. We discover the tourist visiting patterns deeper by combining perception tone and cognitive concern results using network analysis to map out the destinations’ popularity, interconnectivity, and major cognitive perception. Most of the tourists disclose positive expressions and give their concerns about Bali’s natural attractions. They feel best for the social setting and environment aspect, and worst for the accessibility. Sacred Monkey Forest Sanctuary is the most favorite destination and a potential point of a visit to other destinations. This research provides insight into the global perception of Bali’s topmost destinations for government and other tourism stakeholders to support the development and improvement of Bali’s tourism.
Pretrained language models posses an ability to learn the structural representation of a natural language by processing unstructured textual data. However, the current language model design lacks the ability to learn factual knowledge from knowledge graphs. Several attempts have been made to address this issue, such as the development of KEPLER. KEPLER combines the BERT language model and TransE knowledge embedding method to achieve a language model that can incorporate knowledge graphs as training data. Unfortunately, such knowledge enhanced language model is not yet available for the Indonesian language. In this experiment, we propose IndoKEPLER: a language model trained usingWikipedia Bahasa Indonesia andWikidata. We also create a new knowledge probing benchmark named IndoLAMA to test the ability of a language model to recall factual knowledge. The benchmark is based on LAMA, which is designed to test the suitability of our language model to be used as a knowledge base. IndoLAMA tests a language model by giving cloze style question and compare the prediction of the model to the factually correct answer. This experiment shows that IndoKEPLER increases the ability of a normal DistilBERT model to recall factual knowledge by 0.8%. Moreover, the most significant increase happens when dealing with many-to-one relationships, where IndoKEPLER outperforms it’s original text encoder model by 3%.
Network of vehicles using Internet of Things (IoT) frameworks have efficient characteristics of modern intelligent transportation system with a few challenges in vehicular ad-hoc networks (VANETs). However, its security framework is required to manage trust management by preserving user privacy. Wireless mobile communication (5G) system is regarded as an outstanding technology that provide ultra-reliable with limited latency wireless communication services. By extension, integrating Software Defined Network (SDN) with 5G-VANET enhances global information gathering and network control. Therefore, real-time IoT application for monitoring transport services is efficiently supported. These ensures vehicular security on this framework. This paper provides a technical solution to a self-confidential framework for a smart transport system. This process exploiting IoT for vehicle communication by incorporating SDN and 5G technology. Due to some features of blockchain, this framework has been implemented to provide various alternative support for vehicular smart services. This involves real-time access to cloud to stream video information and protection management to vehicular network. The implemented framework presents a promising technique and reliable vehicular IoT environment while ensuring user privacy. Results of simulation presents that vehicular nodes/messages (malicious) and overhead is detected and the impact on network performance are satisfactory when deployed in large-scale network scenarios.
The use of data visualization in defense is very important. In the application of data visualization can use GIS applications. GIS can help a country in maintaining the integrity of the country. This study aims to show how data visualization using GIS is used in defense. Methodology This study uses a qualitative research methodology. The results of this study are that by knowing visualization data assisted by GIS applications, you can find out the Hazard and Vulnerability of the terrain in the country. With data visualization, it can also provide the supply chain needed in the defense industry as well as in times of war. And everyone who needs this visualization data can better plan the network plan and urban plan that will be used.
The left ventricular of ejection fraction is one of the most important metric of cardiac function. It is used by cardiologist to identify patients who are eligible for life-prolonging therapies. However, the assessment of ejection fraction suffers from inter-observer variability. To overcome this challenge, we propose a deep learning approach, based on hierarchical vision Transformers, to estimate the ejection fraction from echocardiogram videos. The proposed method can estimate ejection fraction without the need for left ventrice segmentation first, make it more efficient than other methods. We evaluated our method on EchoNet-Dynamic dataset resulting 5.59, 7.59 and 0.59 for MAE, RMSE and R2 respectivelly. This results are better compared to the state-of-the-art method, Ultrasound Video Transformer (UVT). The source code is available on https://github.com/lhfazry/UltraSwin.
