7 research outputs found
Pengukuran Kemiripan Berbasis Leksikal dan Semantik untuk Perangkingan Dokumen Berbahasa Arab
Get PDFPerangkingan dokumen merupakan salah satu topik dalam sistem temu kembali informasi. Dalam menghasilkan dokumen yang relevan, pengukuran kemiripan antara query dan dokumen menjadi faktor penting terhadap dokumen yang dirangking. Pengukuran kemiripan dapat dihitung berdasarkan bobot kata antara query dan dokumen. Namun, pengukuran kemiripan menggunakan bobot kata dimungkinkan adanya lafal kata yang berbeda tetapi memiliki makna kata yang sama. Selain itu, hasil dokumen pencarian suatu teks berbahasa Arab dipengaruhi oleh beragamnya penguasaan pengguna dalam memahami bahasa Arab. Oleh sebab itu, penelitian ini mengembangkan pengukuran kemiripan secara leksikal untuk mengatasi lafal kata dan pengukuran kemiripan secara semantik untuk mengatasi makna kata. Penggabungan perhitungan kemiripan leksikal dan semantik dihitung berdasarkan bobot kata (leksikal) dan digabungkan dengan word embedding (semantik). Berdasarkan hasil uji coba pada 2900 kitab berbahasa Arab, metode usulan memiliki rata-rata recall, precision, dan f-measure tertinggi daripada metode lainnya sebesar 72.42%, 65.83%, 64.2% pada all query, kemudian 73.2%, 63.15%, 63.1% pada short query, serta 71.31%, 69.86%, 65.7% pada long query. Short query adalah query dengan frekuensi sebanyak 1-2 kata sedangkan long query adalah query dengan frekuensi lebih dari 2 kata.
==================================================================================================================================
Document ranking is one of the topics in the information retrieval. In producing relevant documents, measuring the similarity between the query and the document becomes an important factor for the ranked documents. The similarity measurement can be calculated based on the term weight between the query and the document. However, the calculation of similarity based on the term weights has the possibility of differences in the calculation of weights on terms that are written differently but have the same word meaning. In addition, the results of searching documents for an Arabic text are influenced by the variety of user mastery in understanding Arabic. Therefore, a lexical similarity measurement is developed to overcome word pronunciation and a semantic similarity measurement is developed to overcome word meaning. The combination of lexical and semantic similarity calculations is calculated based on term weights (lexical) and combined with word embedding (semantic). Based on the results of evaluation on 2900 Arabic books, this research method has the highest average recall, precision, and f-measure compared to other methods of 72.42%, 65.83%, 64.2% for all queries, then 73.2%, 63.15%, 63.1% for short queries, and also 71.31%, 69.86%, 65.7% on long queries. A short query is a query with a word frequency of 1-2 words, while a long query is a query with frequency of more than 2 words
Improvement of Cluster Importance Algorithm with Sentence Position for News Summarization
No full textEnhancing YOLOv5s with Attention Mechanisms for Object Detection in Complex Backgrounds Environment
Get PDFEnhancing performance for object detection in complex environments is essential for real-world applications that represent complexities, such as stacking objects in the same location or environment. Models for detecting objects developed to this day still have difficulties in detecting objects with environments that have complex backgrounds. The reason is that the model often experiences a decrease in accuracy when the object to be detected is occlusion by other objects and is small in size. Therefore, in this study, a model improvement method was carried out in detecting objects in a complex environment. The algorithm used in this study is YOLOv5s. Optimization is carried out by adding a CBAM (Convolutional Block Attention Module) attention mechanism layer which is integrated with the C3 layer (C3CBAM) in the backbone of the YOLOv5s model architecture. In addition, a P2 feature map is also added to the architecture head. The optimization results carried out were quite satisfactory, namely there was an increase in the precision value by 1.6 %, at [email protected] an increase of 1.4 %, and also mAP@50-95 increased by 0.1%. This proves that the enhancement method applied to YOLOv5s in this study can improve the performance of the model. However, with the addition of the attention mechanism layer, it turns out that it can increase the computational load. Therefore, for future research, a method can be applied to reduce computing load, one of the methods is knowledge distillation
Sertifikasi Indikasi Geografis Kopi: Pendekatan Studi Bibliometrik
Get PDFGeographical indication certification of coffee is growing due to the demand for "specialty" coffee based on geographic origin. According to coffee lovers, the taste of coffee varies, depending on where the coffee is produced. As demand for geographic indication certified coffee increased, the research that related to this topic also developed. To find out the extent of research that has been carried out regarding the topic, this bibliometic research was conducted. The purpose of this research is to map research on geographical indication certification of coffee based on keywords and research titles so that gaps and novelties related to the topic are obtained. The type of research used is a quantitative descriptive method. The analysis used is bibliometric analysis with the Vosviewers software tool. Based on the analysis results, it was obtained from the network visualization mapping that there were 6 clusters originating from 3032 terms with 50 keywords that appeared at least 10 times. In research on geographical indication certification, the 3 most keywords that appeared were coffee, Indonesia, geographical indications. From the mapping visualization results, it is known that the latest research topics studied are signs, Arabica coffee, coffee farmers, factors, indicators, Indonesian coffee, Robusta coffee, quality, and West Java. The average publication related to these items was published in 2020-2021. From the results of the visualization mapping, the density of research topics related to coffee items in cluster 1, Indonesian items in cluster 4, and geographical indication items in cluster 2 have been widely researched, while topics in clusters 3, 5 and 6 have not yet been widely researched. These items are signs, coffee farmers, indicators, value, quality, coffee farmers, Robusta coffee, quality, and West Java.Sertifikasi indikasi geografis kopi berkembang karena adanya permintaan kopi “speciality” berdasarkan asal geografis. Citarasa kopi menurut kalangan pecinta kopi berbeda, tergantung dimana asal kopi tersebut dihasilkan. Seiring dengan berkembangnya permintaan akan kopi sertifikasi indikasi geografis, penelitian terkait topik ini juga banyak dilakukan. Untuk mengetahui sejauh mana perkembangan penelitian-penelitian maka dilakukanlah penelitian bibliometik ini. Adapun tujuan penelitian ini adalah untuk memetakan penelitian-penelitian sertifikasi indikasi geografis kopi berdasarkan kata kunci dan judul penelitian sehingga diperoleh gap dan novelty terkait topik. Jenis penelitian yang digunakan adalah metode deskriptif kuantitatif. Analisis yang digunakan adalah analisis bibliometrik dengan alat bantu software Vosviewers. Berdasarkan hasil analisis diperoleh pemetaan visualisasi jaringan terdapat 6 cluster yang berasal dari 3032 term dengan 50 kata kunci yang minimal kemunculannya 10 kali. Dalam penelitian sertifikasi indikasi geografis 3 kata kunci terbanyak yang muncul adalah kopi, Indonesia, indikasi geografis. Dari hasil pemetaan visualisasi hamparan diketahui topik penelitian terbaru yang diteliti yaitu tanda, kopi arabika, petani kopi, faktor, indikator, kopi Indonesia, kopi robusta, kualitas, dan Jawa Barat. Rata-rata publikasi yang berkaitan dengan item-item tersebut dipublikasi pada tahun 2020-2021. Dari hasil pemetaan visualisasi kepadatan, topik penelitian yang berkaitan dengan item kopi pada cluster 1, item Indonesia pada cluster 4, dan item indikasi geografis pada cluster 2 sudah banyak diteliti sedangkan topik-topik yang berada pada cluster 3, 5 dan 6 masih belum banyak diteliti. Adapun item-item tersebut adalah tanda, petani kopi, indikator, nilai, kualitas, petani kopi, kopi robusta, kualitas, dan Jawa Barat
MULTI-CLASS REGION MERGING FOR INTERACTIVE IMAGE SEGMENTATION USING HIERARCHICAL CLUSTERING ANALYSIS
Get PDFIn interactive image segmentation, distance calculation between regions and sequence of region merging is being an important thing that needs to be considered to obtain accurate segmentation results. Region merging without regard to label in Hierarchical Clustering Analysis causes the possibility of two different labels merged into a cluster and resulting errors in segmentation. This study proposes a new multi-class region merging strategy for interactive image segmentation using the Hierarchical Clustering Analysis. Marking is given to regions that are considered as objects and background, which are then referred as classes. A different label for each class is given to prevent any classes with different label merged into a cluster. Based on experiment, the mean value of ME and RAE for the results of segmentation using the proposed method are 0.035 and 0.083, respectively. Experimental results show that giving the label on each class is effectively used in multi-class region merging
Improvement of Cluster Importance Algorithm with Sentence Position for News Summarization
Get PDFText summarization is one of the ways to reduce large document dimension to obtain important information from the document. News is one of information which usually has several sub-topics from a topic. In order to get the main information from a topic as fast as possible, multi-document summarization is the solution, but sometimes it can create redundancy. In this study, we used cluster importance algorithm by considering sentence position to overcome the redundancy. Stages of cluster importance algorithm are sentence clustering, cluster ordering, and selection of sentence representative which will be explained in the subsections below. The contribution of this research was to add the position of sentence in the selection phase of representative sentence. For evaluation, we used 30 topics of Indonesian news tested by using ROUGE-1, there were 2 news topics that had different ROUGE-1 score between using cluster importance algorithm by considering sentence position and using cluster importance. However, those 2 news topics which used cluster importance by considering sentence position have a greater score of Rouge-1 than the one which only used cluster importance. The use of sentence position had an effect on the order of sentence on each topic, but there were only 2 news topics that affected the outcome of the summary
Pembuatan alat pemurni air laut skala besar untuk memenuhi kebutuhan air bersih di SMA Negeri 1 Rupat daerah pesisir Provinsi Riau
Get PDFRupat Island located at a coastal area of Riau Province, generally coastal areas are identical with the availability of clean water. The low availability of clean water has a negative impact on all sectors. The water used by students and teachers at SMA Negeri 1 Rupat comes from a ring well which tastes salty and brackish. Problems in the health sector are the availability of clean water that is suitable for consumption, education is insufficient equipment for Physics Experiments, in the economic field, family spending increases due to the availability of bottled water, and the field of community empowerment has not maximized the use of natural resources. The solution for this aspect is technological innovation that is able to meet the need for clean water in the long term by utilizing existing potential, one of which is the manufacture of a sea water purifier, a giant measuring 4x4 meter. The tool's fundamental working principle is to evaporate the air first and then collect the condensed water to separate the mineral salts and contaminants from the seawater. The author has made a seawater purifier measuring 48x38 cm as a medium for high school physics learning in 2021 with a TKT 5
