Karakteristik teks yang tidak terstruktur menjadi tantangan dalam ekstraksi fitur pada bidang pemrosesan teks. Penelitian ini bertujuan untuk membandingkan kinerja dari word embedding seperti Word2Vec, GloVe dan FastText dan diklasifikasikan dengan algoritma Convolutional Neural Network. Ketiga metode ini dipilih karena dapat menangkap makna semantik, sintatik, dan urutan bahkan konteks di sekitar kata jika dibandingkan dengan feature engineering tradisional seperti Bag of Words. Proses word embedding dari metode tersebut akan dibandingkan kinerjanya pada klasifikasi berita dari dataset 20 newsgroup dan Reuters Newswire. Evaluasi kinerja diukur menggunakan F-measure. Performa terbaik menunjukkan FastText unggul dibanding dua metode word embedding lainnya dengan nilai F-Measure sebesar 0.979 untuk dataset 20 Newsgroup dan 0.715 untuk Reuters. Namun, perbedaan kinerja yang tidak begitu signifikan antar ketiga word embedding tersebut menunjukkan bahwa ketiga word embedding ini memiliki kinerja yang kompetitif. Penggunaannya sangat bergantung pada dataset yang digunakan dan permasalahan yang ingin diselesaikan.Kata kunci: word embedding, word2vec, glove, fasttext, klasfikasi teks, convolutional neural network, cnn.
The remote control system on electrical equipment in the room can be fulfilled through the internet as an IoT (Internet of Things) implementation. All devices managed from one interface, so home appliances management delivered quickly and conveniently. The main contribution in this research is IP based controlling for rooms with control lights and vertical curtains, and also the temperature of the air conditioner (AC) with IoT Technology. The used hardware is Raspberry Pi 3 as a server, Relay, motor stepper, IR led Transmitter, and temperature sensor DS18B20. For implementation, an android application is built by MIT App Inventor 2. The results show that all features function correctly, but each device responds with a different delay value. Delay time response of a lamp, vertical blind, and AC is up to 1.5 sec, 2.1 sec, and 1.6 sec, respectively.electrical appliances, IoT, controlling system, smart room
The wind carries moisture into an atmosphere and hot or cold air into a climate, affecting weather patterns. Knowing where the wind is coming from gives essential insight into what kind of temperatures are to be expected. However, the wind is affected by spatial and temporal variabilities, thus making it difficult to predict. This study focuses on finding data associations from the weather station installed at Hasanuddin University Campus based on internet of things (IoT) using Raspberry Pi as a gateway that associated all the meteorological data from sensors. The generation of association rules compares the Apriori and FP-growth algorithms to determine relations among itemsets. The results show that high humidity and warm temperature tend to associate with a westerly wind and occur at night. In contrast, conditions with less humid and moderate temperatures tend to have southerly and southeasterly wind.
Sinyal suara terkadang masih mengandung noise pada saat dilakukan proses pengolahan. Filter dilakukan untuk menyaring sinyal-sinyal suara yang tidak dibutuhkan ataupun dianggap mengganggu (noise). Penelitian ini bertujuan untuk melakukan preprocessing sinyal suara dengan membandingkan kinerja filter Infinite Impulse Respon (IIR) pada desain filter Butterworth berdasarkan frekuensi yang dilewatkan. Tujuan yang ingin dicapai dari tahap preprocessing yaitu mengolah suara agar dapat diambil karakteristik atau cirinya. Sinyal suara sampel yang akan difilter yaitu perekaman suara laki-laki pada frekuensi sampling 16000Hz. Pada penelitian ini filter Infinite Impulse Respon (IIR) digunakan dengan respon filter Butterworth. Validasi dilakukan dengan menghitung Signal Noise to Ratio (SNR) dari masing-masing jenis filter berdasarkan frekuensi yang dilewatkan. Hasil yang diperoleh berdasarkan nilai SNR tertinggi yaitu 29,9321 dB pada orde filter 6. Nilai dari SNR ini juga menunjukkan bahwa BPF lebih baik dibandingkan dengan LPF dan HPF.
Abstract-In this study, speech to text system for homophone phrases in Indonesian was designed using an extraction method which featured Mel Frequency Cepstral Coefficient (MFCC). Feature extraction results were classified by comparing the two classifiers of Backpropagation Neural Network (BPNN) and KNearest Neigbour (KNN). The input data used were the recordings of each of 3 male and female respondents. The recording process was conducted for 5 seconds at a sampling frequency of 16 kHz and at channel mono. Classification results with test data to BPNN showed accuracy rates of 96.67% and 90% respectively for male and female respondents. Moreover, the level of accuracy obtained on KNN amounted to 83.33% for males and 73.33% for females.
To finish study, students are requested to submit final projects. In some universities, the final projects are not necessary to be submitted for publication. The final project reports are stored in a local database. As the number of final projects is growing in the local database, similar contents may exist among the documents. The commercial tools cannot be used to detect the content similarity since the documents are not published. This paper proposed a system to detect content similarity in documents that are stored in a local database. Considering the number of stored documents, this similar content detection system implements two step processes. First, clustering documents to find most related documents. Second, finding content similarity among the selected documents. The experiment results show that the system is successfully clustering documents and detecting content similarity by implementing TF-IDF and Cosine Similarity algorithms. This system is limited to proceed documents that are written in Bahasa.
This study was aiming at helping visually impaired people to detect and estimate the fire distance. Blind people had difficulty knowing the existence of fire at a safe distance; hence the possibility of burning could occur. The color models and blob analysis methods were used to detect the presence of fire in the blind path. Before the fire detection stage, the cascade of the HSV and RGB color models was applied to segment the reddish fire color. The size and shape of a dynamic fire were the parameters used in this paper to distinguish fire from non-fire objects. Changes in the area of the fire object obtained at the Blob analysis stage per 10 frames were the main contributions and novelty in this paper. After the fire is detected, the calculation of the fire distance to a blind person was completed using a pinhole model. This research used 35 data videos with a resolution of 480x640 pixels. The results showed that the fire detection system and the distance estimation achieved an accuracy of 88.86% and the MSE of 0.0358, respectively.
The development of modern retail business is gradually getting faster, increasing the level of competition among retailers. The retailers are changing their business strategies to acquire new customers, maintain customer loyalty and improve customer service. One way to indicate the good market performance of a retail shop is to know the number and details of visitors based on gender. The number of visitors in retail shops can be observed by installing CCTV camera. In the existing shop system, CCTV used just for monitoring activities in the retail shop. Therefore, the data from the CCTV camera can be used to calculate the number of visitors based on gender automatically by utilizing computer vision. This study designed a system to count the number of female visitors through video data. The data acquisition process involved 89 people, 41 women and 48 men. The data are stored in .AVI video file format with a resolution of 7201280 pixels. This system can be divided into three main stages, which are face detection using the Viola-Jones method, feature extraction using Gabor Filter 2D and classification using Support Vector Machine (SVM) method. The result of the study showed that the system can count the number of female and non-female visitors with an accuracy rate of 96.52%. The system performance will be improved by using another feature extraction and classification methods.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.