1 January 2017 in Procedia Computer Science
The quality of a tourism products can be assessed using several aspects or objects due to their unique characteristics. The information related to the object can be extracted using object-based opinion mining. Based on the previous research, the implementation of Natural Language Processing (NLP) rules on object-based opinion mining for determining the orientation of the semantic objects showed good result. However, the performance of the objects extraction should be improved. In this study, researchers apply a filter on the objects’ extraction process of the hotel and restaurant review data. The utilization of data filter in object-based opinion mining succeeded in obtaining better objects’ extraction result due to the utilization of filter that eliminate the unrelated object. The application of filter in the process of objects extraction improve the precision of frequent object approach from 45.7% to 64.49% on of the hotel review and from 44.82% to 64.61% on the restaurant review. For frequent and infrequent approach, the precision was increased from 22.33% to 63.02% on the hotel review and from 21.6% to 65.4% on the restaurant review. For overall extracted object, the usage of filter got better result compared to non-filter classification process. The filtered object approach gave 56.85% accuracy, 60.91% precision, and 79.93% recall on the hotel review, and got 58.85% accuracy, 63.26% precision, and 84.14% recall on the restaurant review.
22 September 2024 in Annals of Operations Research
19 September 2024 in IEEE Access
The timetabling problem is a class of optimization problems applicable to various domains, such as education, healthcare, sports, and transportation. Characterized by its computational complexity, the problem is categorized as NP-Hard, suggesting heuristic methodologies as common solutions. Throughout academic research, a variety of algorithms and benchmark datasets has been proposed, contributing to an expanding collection of knowledge in the field. However, most of the reported successful algorithms are compared with a limited number of prior studies. This makes it difficult for researchers to ascertain the efficiency of the proposed algorithms. This lack of comparison extends to the difficulty of assessing generic algorithms effective for solving timetabling problems. Addressing this gap, our study conducts a comprehensive survey of 64 papers published between 2012 and 2022, aiming to provide the state of the art of timetabling problems benchmarks and algorithms as well as a critical analysis and comparative evaluations. Our survey provides three contributions. First, this paper surveys the state of the art from problem instances and algorithms in timetabling problems. Second, this paper provides comprehensive algorithm performance on timetabling problems. Lastly, we provide a survey on the generic algorithm for cross-domain timetabling. Our findings indicate that single solution-based metaheuristics approaches, particularly the Simulated Annealing algorithm, showthe most promise in addressing the timetabling problem. Moreover, we found only a few studies that successfully demonstrate consistent performance across multiple benchmarks and distinct problem sets. This survey serves as a foundational element for future research towards effective algorithmic strategies for single and cross-domain timetabling problems.
31 July 2022 in MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer
Pengenalan objek bawah laut dapat dilakukan berdasarkan pola hamburan SONAR, seperti untuk deteksi ranjau dan deteksi batu yang terletak di dasar laut. Kesulitan yang dihadapi pada pengenalan objek bawah laut antara lain adalah pemilihan metode ekstraksi fitur, adanya rotasi objek yang menghasilkan pola hamburan yang berbeda, lingkungan atau latar belakang bervariasi, dan kemampuan pengklasifikasi yang berbeda untuk lingkungan yang lebih kompleks. Pada penelitian ini, kami menggunakan deep learning neural network untuk meningkatkan kinerja klasifikasi dua buah objek bawah laut. Secara khusus, dibandingkan jumlah neuron pada lapisan tersembunyi dan fungsi aktivasi yang dapat menghasilkan kinerja yang lebih tinggi dari penelitian sebelumnya. Pada penelitian sebelumnya, proses klasifikasi dilakukan dengan menggunakan neural network dengan 12 buah lapisan tersembunyi, dan menghasilkan akurasi maksimal sebesar 90.4%. Dilakukan percobaan pada struktur jaringan syaraf tiruan berupa multilayer perceptron dengan 2 buah lapisan tersembunyi dan 7 macam fungsi aktivasi. Dari percobaan yang dilakukan diperolehbahwa deep learning neural network memberikan rata-rata akurasi terbaik sebesar 85,9% dengan akurasi maksimal sebesar 96,15% lebih baik dibandingkan hasil penelitian sebelumnya. Akurasi terbaik tersebut diperoleh dengan memanfaatkan jumlah neuron pada lapisan tersembunyi sebanyak 140 buah, dan fungsi aktivasi reLU untuk lapisan tersembunyi fungsi aktivasi Linear untuk lapisan output.