24 research outputs found
Simulasi Antrian Satu Channel Dengan Tipe Kedatangan Berkelompok
No full textMasalah antrian tidak hanya terjadi dalam kegiatan sehari – hari namun juga dapat terjadi pada suatu sistem komputer. Antrian yang akan dibahas memiliki sebuah server dengan satu garis antrian yang melayani unit dalam antrian satu per satu dengan tipe kedatangan berkelompok. Pola kedatangan pada antrian ini berdistribusi Poisson dan pola pelayanan berdistribusi Eksponensial dengan disiplin antrian FIFO ( First In First Out ). Untuk mengamati perilaku sistem antrian digunakan simulasi yang akan dijalankan dengan memberikan input yang berbeda-beda dan akan mempengaruhi output sistem. Dari hasil simulasi diharapkan dapat diketahui karakteristik sistem antrian terutama probabilitas kesibukan server sehingga dapat dijadikan landasan untuk pengambilan keputusan terhadap sistem antrian yang diamati
Information Extraction from Web as Knowledge Resources for Indonesian Question Answering System
No full textSimulasi Antrian Satu Channel Dengan Tipe Kedatangan Berkelompok
No full textSimulasi Antrian Satu Channel Dengan Tipe Kedatangan Berkelompo
STEGANOGRAFI CITRA RGB DENGAN PENGACAKAN BLOK DAN PRAPROSES ENKRIPSI MENGGUNAKAN ALGORITMA RIJNDAEL 128 BIT
No full textAbstract— Sending message via internet is too dangerous, because the other wants to know the body of message. This time, cryptography technique has been used to guard the message, but this technique invites the other to attack that encrypted message. So that, steganography technique is added in order to make the encrypted message protected and doesn’t invite the attack. The steganography embedding technique that used in this research is LSB technique with randomize embedding block which using the randomize number result of PRNG. Cryptography algorithm which used in this research is 128 bit rijndael algorithm. After all those techniques have been implemented into program and the embedded have been done into 24 bit(RGB) bitmap image, stego-image publish into some forums to be attacked and try to be attacked by some steganalysis tools, and until March 17th, 2012 from 4 forums and 4 steganalysis tools, the result is none of forums or steganalysis tools can extract the message. It can be concluded that the stego-image is strong from steganalysis attacks until March 17th, 2012 and the techniques that used in this research are success to make difficult another in analyzing and extracting the message. For the next research, writer hope that the researcher can use the other image format, and then the software will be added with image expansion feature. For addition, the software is more developed in order to make the stego-image strong when the image attacked by image resize attack, image cropping attack and many so on. Keywords: Cryptography, Rijndael, Steganography, Least Significant Bit, Pseudo Random Number Generator, stego- image, steganalysis. Abstrak-- Pengiriman pesan melalui media internet bisa berbahaya, karena banyak pihak-pihak yang tidak berkepentingan ingin mengetahui isi pesannya. Pada saat ini sudah digunakan teknik kriptografi untuk mengamankan pesan, tetapi justru teknik ini yang mengundang pihak yang tidak berkepentingan untuk menyerang pesan terenkripsi tersebut. Oleh karena itu, ditambahkanlah teknik steganografi agar pesan yang terenkripsi tersebut tersamar dan tidak mengundang serangan. Teknik penyisipan yang digunakan adalah teknik LSB dengan pengacakan blok penyisipan berdasarkan hasil pembangkitan bilangan acak menggunakan PRNG. Algoritma kriptografi yang digunakan adalah algoritma rijndael 128 bit. Setelah diimplementasi ke dalam program dan dilakukan penyisipan ke dalam citra bitmap 24 bit(RGB), stego-image yang dihasilkan dilempar ke beberapa forum untuk diserang, serta diuji menggunakan beberapa tools steganalisis, hasilnya sampai tanggal 17 Maret 2012, dari 4 forum dan 4 tools steganalisis, tidak ada satupun forum ataupun tools steganalisis yang berhasil mengekstrak pesan dari stego-image tersebut. Dari hasil pengujian ini dapat disimpulkan bahwa, stego-image yang dihasilkan dinyatakan tangguh dari serangan steganalisis sampai tanggal 17 Maret 2012 dan teknik-teknik yang digunakan dalam penelitian ini berhasil mempersulit pihak lain untuk menganalisis dan mengekstrak pesannya. Diharapkan untuk penelitian selanjutnya bisa menggunakan format citra yang lain, kemudian perangkat lunak ditambahkan fitur ekspansi wadah dan dikembangkan lagi agar stego-image tetap tangguh ketika mendapatkan serangan resize image, cropping image dan lain-lain. Kata kunci : Kriptografi, Rijndael, Steganografi, Least Significant Bit, Pseudo Random Number generator, stego- image, steganalisis
NL2SQL For Chatbot with Semantic Parsing Using Rule-Based Methods
No full textStructured Query Language (SQL) is a command language that allows users to access database information. Ordinary people generally donot know how to make queries with SQL to a database. The chatbot is acomputer program developed to interact with its users via text or voice. In this study, chatbots were developed to help and facilitate users intheNatural Language to Structured Query Language (NL2SQL) process tosearch for information in an Academic Information Systemdatabasewith semantic parsing using a rule-based method that accepts input inthe form of interrogative sentences or order. In the Natural Language toStructured Query Language (NL2SQL) process several problems arise, namely input problems with unique parameters for the knowledge base, and slow searching or translation processes, which make Natural Language to Structured Query Language (NL2SQL) inef icient, problems This problem will be solved using a semantic parsingapproach using a rule-based method that is proven to be ef icient insolving issues such as the Natural Language to Structured QueryLanguage (NL2SQL) process. The results showed that the semanticparsing approach using the rule-based method succeeded in obtainingan accuracy rate of 96.72% using 122 test data in the formof questionsentences or command data about the Academic Information Systemof the Department of Informatics Engineering, Sriwijaya University inIndonesian, and an average execution time of 50.68 milliseconds. seconds or 0.05 seconds
Bully Comments Classification on TikTok Using Support Vector Machine and Chi-Square Feature Selection
No full textTikTok has been named the world’s most popular social media platform. The high level of TikTok use makes it easier for an irresponsible user to do unethical things such as spreading hateful comments on someone’s account. TikTok developers can prevent bullying by using policies such as word detection and filtering features that indicate comments fall under the category of bullying or non-bullying comments. Therefore, we conducted this study to classify bullying comments using Machine Learning methods for convenience purposes on TikTok usage, a method that we used in this research is the SVM method to classify the data and Chi-Square as the feature selection. Tests were carried out using the Linear, Polynomial, and RBF kernel functions with the C parameter, namely 0,1, 1, and 10 for each kernel. The results of this research show that the Support Vector Machine method with Chi-Square Feature Selection has a better performance. This was proven by the increased accuracy in RBF kernel C=0,1 which was 0,2
Cat Breeds Classification Using Convolutional Neural Network For Multi-Object Image
No full textCat is one of the most popular pets. There are many cat breeds with unique characteristic and treatment for each breed. A cat owner can have more than one cat, either the same breed or different breeds. But not all cat owners know the breeds of their cats. Computers can be trained to recognized cat breeds, but there are many challenges for computers because it limited by how much they have been trained and programmed. In recent years, a lot of research about image classification has been done before and got various result, but most of the data used in previous research were single object images. Therefore, this study of cat breeds classification would be conducted with Convolutional Neural Network (CNN) in the Multi-Object images. This method was chosen because it had good classification results in the previous studies. This study used 5 breeds of cats with every breed having 200-3200 images for training. The test results were measured using confusion matrix, obtaining the precision, recall, f1 score and accuracy of 100% on multi-object images with 2 objects and 3 objects. On images with 4 objects achieved the precision, recall, f1 score and accuracy value of 89%, 87%, 87% and 95%. While the value of precision, recall, f1 score and accuracy on images with 5 objects get 87%, 86%, 86% and 94%, respectively
Fisheries Harvest Prediction using Genetic Algorithm Optimized of Gated Recurrent Unit
No full textIndonesia is a maritime country with most of the population living near water areas. Water products are a common commodity often consumed cheaply, and food is therefore one of the primary human needs. Fishery harvest predictions are needed to control prices, prepare seeds, and ensure stable sales and consumption. The reason for choosing GRU for this prediction is that classical methods, commonly used in econometrics or time series analysis, were previously prevalent. GRU requires fewer operations than LSTM. Instead of training with an optimization algorithm relying on backpropagation and gradients, metaheuristic optimization in the form of a GA is used. GA does not require gradient information and is expected to avoid local optima. The total average MSE obtained is 9.55%
