Abstract:Automatic detection of hate speech and abusive language is crucial for combating online toxicity. This study explores Gaussian Naive Bayes for multi-label classification of hate speech on Indonesian Twitter, including target,…
rget, category, and level. We combined TF-IDF features with contextual BERT embeddings. The model achieved balanced performance for general hate speech and good non-abusive language detection. However, it exhibited limitations with imbalanced data and specific hate speech types. The classifier consistently favored the majority class (non-hateful/non-abusive) across labels, particularly struggling with HS_Gender, HS_Physical, etc. This suggests difficulty detecting less frequent but potentially severe hate speech, likely due to limited training data. Overall accuracy and F1-scores confirm that while Gaussian Naive Bayes is efficient, it lacks robustness for nuanced multi-label classification with imbalanced datasets. This necessitates exploring alternative approaches for effectively detecting specific and less frequent hate speech.
Abstract:An English folksong entitled The Wellerman became popular in Indonesia recently. Therefore, the accuracy and acceptability in folksong were important to be investigated. This paper reported the accuracy and acceptability…
of English into Indonesian translation in a folksong entitled The Wellerman. This research applied descriptive qualitative method. In addition, this research combined with purposive sampling technique. Nababan translation quality assessment framework was used to investigate the accuracy and acceptability of the folksong. The whole The Wellerman’s lyrics consist of 28 lines which was then assessed by the raters in terms of its accuracy and acceptability. The raters were 2 translators and 1 writer. The raters were purposely chosen to conduct reliable results of the accuracy and acceptability. This study found the average score of the accuracy was 2. 35 with 198 total scores and the average score of the acceptability was 2.48 with 209 total scores. The result indicates that the translation was relatively accurate and acceptable for the raters. However, it needs further improvement to maintain the context of the source text (ST) in the target text (TT) to reach an accuracy and acceptability in The Wellerman folksong. This study implied that it is crucial to maintain the imagery, narration, and the context of folksong translation.
Abstract:The demand for electrical energy continues to rise with the progression of time. This growth must be matched by a reliable and cost-effective supply of electricity, requiring power systems that are both dependable and economical.…
onomical. Since the amount of electricity consumed by users cannot be precisely predicted, balancing generation with consumption necessitates accurate electrical load forecasting. This study focuses on load forecasting using the Adaptive Neuro-Fuzzy Inference System (ANFIS) method. The forecast developed targets daily peak loads, which fall under short-term load forecasting. The data used for this forecasting consists of historical daily peak loads from January 1, 2017, to June 9, 2022. The forecasting process involves parameters such as radius, squash factor, accept ratio, reject ratio, and epoch. The forecast accuracy is evaluated using the Mean Absolute Percentage Error (MAPE) metric. The results are then compared with PLN’s load forecasting, which employs the load coefficient method. The ANFIS-based forecasting achieved a MAPE of 1.879%, using networks Jaringan_24 and Jaringan_25. This MAPE value is slightly lower than PLN’s load forecasting MAPE of 1.917%, indicating better accuracy by the ANFIS method.
Abstract:This study aims to examine the authenticity of Sahih Bukhari Hadith No. 2112 regarding the right of khiyar through a comprehensive takhrij method to ensure the validity of the evidence in muamalah policy. Additionally, this…
his study examines the mechanism of transmitting legal texts without editorial changes and compares the ijtihad of the four schools of jurisprudence regarding time limits to provide consumer protection solutions in the digital age. The methodology employed is normative legal research using a descriptive-analytical qualitative approach through library research. Data collection techniques involved cataloging hadiths on khiyar from the Kutubus Sittah, identifying the structure of the isnad, and analyzing key vocabulary (mufradat). Data analysis was conducted through stages of isnad criticism to assess the quality of the narrators, systematic analysis of the matn, comparative analysis across schools of thought, and the synchronization of traditional principles with modern economic realities. The research results indicate that the hadiths on khiyar possess exceptional chain of transmission quality within the Silsilah adz-Dzahab tradition, ensuring the text’s accuracy free from distortion over fourteen centuries. Regarding the time limit for khiyar syarat, differing viewpoints were identified: the Shafi’i school limits it to a maximum of three days, while the Maliki school allows a duration of up to 38 days depending on the type of object. In conclusion, the principle of khiyar remains relevant in the digital economy through the transformation of the order cancellation feature as a manifestation of khiyar majelis, as well as the return policy as an application of khiyar aib and khiyar syarat. The implications of this research emphasize that the ethical values of khiyar can serve as a foundation for regulators in refining consumer protection laws to minimize information asymmetry and ensure full consent (antaradin) in every online transaction
Abstract:This paper discussed the result of Controlled Speaking Practices of Senior Students at SMA. This teaching technique had a big contribution to improve accuracy and fluency in speaking. The study aimed to investigate the strengths…
trengths and weaknesses of Controlled Speaking Practices and it was also to find whether Controlled Speaking Practices is effective in improving the speaking skill of senior students or not. The study took place at SMA TA’MIRIYAH Surabaya. The school carried out Controlled Speaking Practices in conversation class in the first grade of senior high school students. Qualitative descriptive approach and observation were applied in this study. The writer used recording and field notes as the techniques in conducting observation. The writer analyzed the strengths and weaknesses of controlled speaking practice which were according to three aspects: the result of practicing the task, how the teacher conducted the task, and the way students’ talk in the class. From the data analysis, Controlled Speaking Practices also can indicate that this is the right technique to improve the students’ speaking skill. The researcher concluded that Controlled Speaking Practices might be a tool of practice speaking English which can assist the students to achieve good speaking English.
Abstract:Penyakit tanaman merupakan salah satu faktor utama penurunan hasil pertanian pada komoditas tomat dan cabai di Indonesia. Deteksi dini secara manual memerlukan keahlian khusus dan waktu yang lama. Penelitian ini mengusulkan…
kan sistem klasifikasi penyakit tanaman berbasis deep learning menggunakan arsitektur MobileNetV2 dengan pendekatan transfer learning. Dataset PlantVillage yang terdiri dari 20.638 gambar daun dengan 15 kelas digunakan sebagai data pelatihan. Model dievaluasi menggunakan metrik accuracy, precision, recall, dan F1-score. Visualisasi Grad-CAM diterapkan untuk menginterpretasikan area fokus model dalam pengambilan keputusan. Hasil eksperimen menunjukkan akurasi sebesar 89,93% pada data uji dengan rata-rata weighted F1-score sebesar 0,90. Visualisasi Grad-CAM membuktikan model mengidentifikasi area terinfeksi secara akurat. Pengujian pada gambar nyata menunjukkan kemampuan model dalam kondisi dunia nyata.
Abstract:This paper explores the role of Artificial Intelligence (AI) in enhancing decision-making processes within multinational corporations. The primary issue addressed is how AI can be integrated effectively across diverse global…
obal markets, considering factors like regulatory frameworks, cultural diversity, and market dynamics. The research proposes a framework for AI implementation that ensures both operational efficiency and ethical soundness. The study employs a mixed-methods approach, combining qualitative interviews and quantitative surveys from key stakeholders in multinational corporations. Preliminary findings suggest that AI significantly improves decision-making speed and accuracy, particularly in data analysis, market trend prediction, and consumer behavior forecasting. However, challenges remain in adapting AI systems to various cultural and regulatory environments, highlighting the need for customization and local adjustments. This study contributes to understanding how AI can be applied more effectively and ethically across international markets, offering insights for future implementations in diverse business contexts.
Abstract:Kanker payudara merupakan salah satu jenis kanker yang paling umum ditemukan pada wanita dan menjadi penyebab utama kematian akibat kanker di seluruh dunia. Ketepatan dalam diagnosis kanker payudara menjadi sangat krusial…
l untuk penanganan yang tepat. Penelitian ini bertujuan untuk membandingkan performa algoritma Naive Bayes dan Random Forest dalam mengklasifikasikan kanker payudara menggunakan dataset Breast Cancer Wisconsin. Metodologi penelitian dimulai dengan pengumpulan data dari dataset Breast Cancer Wisconsin yang terdiri dari 569 sampel dengan 32 atribut. Proses preprocessing data meliputi konversi data dari format nominal ke binominal untuk atribut diagnosis. Implementasi algoritma menggunakan tools RapidMiner dengan pendekatan cross validation (k=10) untuk evaluasi model yang lebih robust. Performa kedua algoritma dibandingkan menggunakan berbagai metrik evaluasi termasuk accuracy, precision, recall, dan analisis confusion matrix. Hasil penelitian menunjukkan bahwa algoritma Random Forest memberikan performa yang lebih unggul dengan tingkat akurasi 94,91% (±5,06%), precision 95,33%, dan recall 93,90%. Sementara itu, Naive Bayes mencapai akurasi 93,51% (±5,30%), precision 93,68%, dan recall 92,67%. Random Forest juga menunjukkan keunggulan dalam mengurangi false positive, dengan hanya 8 kasus dibandingkan 15 kasus pada Naive Bayes. Analisis confusion matrix menunjukkan bahwa kedua algoritma memiliki kemampuan yang baik dalam mengklasifikasikan kasus kanker payudara, meskipun Random Forest menunjukkan performa yang lebih stabil dan akurat. Kesimpulan dari penelitian ini menunjukkan bahwa kedua algoritma efektif untuk klasifikasi kanker payudara, dengan Random Forest menunjukkan keunggulan dalam hal akurasi dan presisi. Hasil ini dapat menjadi pertimbangan dalam pengembangan sistem pendukung keputusan untuk diagnosis kanker payudara, dimana Random Forest dapat menjadi pilihan utama ketika akurasi menjadi prioritas, sementara Naive Bayes tetap menjadi alternatif yang valid ketika kesederhanaan implementasi dan efisiensi komputasi diperlukan.
Abstract:The use of Artificial Intelligence (AI) technology to visualize the narrative of English-language short stories offers a novel approach to enriching the interpretive and aesthetic dimensions of storytelling. Utilizing BingImage.com…
ngImage.com as an AI-based image generator, textual narratives are transformed into prompts that create visual depictions of each scene. The visuals effectively reflect the mood, emotions, and settings of the story, albeit with some inconsistencies in character depictions across scenes. This exploratory approach examines correlations between narrative and visuals, scene continuity, and aesthetic accuracy. The findings indicate that AI-based visualization can create a more immersive reading experience, enhancing the reader's comprehension through relevant illustrations. This technology also provides creative interpretations that expand the artistic dimensions of short stories. Despite its potential, challenges such as maintaining visual consistency across scenes remain. These findings provide new insights into the integration of digital technology in literary transformation, paving the way for innovative future methodologies.
Abstract:Stunting pada balita merupakan permasalahan kesehatan masyarakat yang serius dan memerlukan penanganan segera melalui deteksi dini yang akurat. CNN (Convolutional Neural Network) merupakan metode yang akurat, efektif dan…
tepat sasaran dalam memberikan hasil yang akurat dan presisi. Implementasi metode CNN dapat mengklasifikasikan status stunting pada balita berdasarkan parameter antropometri dan karakteristik kesehatan. Dataset yang digunakan terdiri dari 6.500 sampel data balita dengan 8 variabel meliputi jenis kelamin, usia, berat lahir, panjang lahir, berat badan, panjang badan, riwayat ASI eksklusif, dan status stunting. Metodologi yang digunakan melibatkan serangkaian tahapan preprocessing data termasuk standardisasi fitur menggunakan StandardScaler, pemisahan data training (80%) dan testing (20%), serta penggunaan arsitektur CNN yang terdiri dari layer konvolusi 1D dengan 32 filter, max pooling, dan dense layer. Model dilatih menggunakan optimizer Adam dengan fungsi loss categorical crossentropy selama 10 epoch dan batch size 32. Evaluasi performa model dilakukan menggunakan berbagai metrik termasuk accuracy, precision, recall, dan F1-score.
Hasil penelitian menunjukkan bahwa model CNN yang dikembangkan mencapai performa yang sangat baik dengan akurasi 90%, precision 89%, recall 92%, dan F1-score 91%. Analisis confusion matrix mengkonfirmasi kemampuan model dalam mengklasifikasikan kedua kelas yaitu pada stunting dan non-stunting secara seimbang. Temuan ini mengindikasikan bahwa implementasi CNN efektif dalam mengidentifikasi status stunting pada balita dan berpotensi menjadi alat bantu yang berharga dalam screening stunting di fasilitas kesehatan.