Abstrak Dalam persaingan ketat saat ini, promosi yang baik dapat memberikan kredibilitas untuk produk baru. Promosi perlu mendapat perhatian lebih dan serius, karena dalam kehidupan sehari-hari timbul produk unggulan, jika tidak mengetahuinya, kemungkinan produk yang ditawarkan kepada konsumen kurang ditanggapi oleh pasar, oleh karena itu perusahaan harus mengupayakan produknya, meyakinkan dan mempengaruhi konsumen untuk menciptakan permintaan akan produk ini. Langkah yang bisa dilakukan oleh perusahaan untuk melakukannya adalah dengan melakukan pemasaran langsung. Peningkatan akurasi prediksi pemasaran langsung dapat dilakukan dengan cara melakukan seleksi terhadap atribut, karena seleksi atribut mengurangi dimensi dari data sehingga operasi algoritma data mining dapat berjalan lebih efektif dan lebih cepat. Dalam penelitian ini akan digunakan metode support vector machine dan akan dilakukan seleksi atribut dengan menggunakan particle swarm optimization untuk prediksi pemasaran langsung. Setelah dilakukan pengujian maka hasil yang didapat adalah support vector machine menghasilkan nilai akurasi sebesar 88,71 %, nilai precision 89,47% dan nilai AUC sebesar 0,896. Kemudian dilakukan seleksi atribut dengan menggunakan particle swarm optimization dimana atribut yang semula berjumlah 16 variabel prediktor terpilih 12 atribut yang digunakan. Hasil menunjukkan nilai akurasi yang lebih tinggi yaitu sebesar 89,38%, nilai precision 89,89% dan nilai AUC sebesar 0,909 dengan nilai akurasi klasifikasi sangat baik (excellent clasiffication). Sehingga dicapai peningkatan akurasi sebesar 0,67 %, dan peningkatan AUC sebesar 0,013. Kata Kunci: Particle Swarm Optimization, Pemasaran Langsung, Seleksi Atribut Abstract In the current intense competition a good promotion can provide credibility for a new product. Promotion needs to get more attention and serious, because in everyday life arise a prime product, if not find out, the possibility of products offered to consumers less responded by the market, therefore the company should strive for its products. , convincing, and influencing consumers to create demand for these products. Steps that can be done by the company to do so is to do direct marketing. Increased accuracy of direct marketing predictions can be done by selecting attributes, because of the selection. Data mining can run more effectively and quickly. In this study the method to be used is. With particle swarm optimization for direct marketing prediction optimization. After testing, the results obtained are support vector engine yield value of 88.71%, precision value 89.47% and AUC value of 0.896. Then the attribute selection is done using particle swarm optimization where the original attribute uses 16 predictor variables selected 12 attributes used. The results showed a higher value of 89.38%, 89.89% accuracy and AUC value of 0.909 with very good fair value (excellent classification). The price increase is 0.67%, and the increase of AUC is 0,013. Keywords: Particle Swarm Optimization, Direct Marketing, Selection Attributes.
At this time information technology is developing very rapidly and requires business actors to be able to follow the developments and advances of the times, especially in the world of technology and information. Inventory of merchandise is a company asset that is one of the assets included in current assets. Merchandise inventories are company assets that are purchased and stored to be resold and make a profit. Recording of merchandise inventory that is still manual, such as data collection of incoming goods, demand for goods, delivery of goods, returns of goods to the preparation of reports will certainly result in the accumulation of goods request notes. The difficulty of data collection of requests for goods from branches to deliver goods, errors in goods requested and sent, the length of time to record the return of goods, errors in calculating the stock of goods and difficulty in obtaining reports when needed are also one of the obstacles in the process of merchandise inventory. Therefore, a web-based inventory information system is needed in order to make it easier for users to manage the inventory process of their merchandise, so that it can simplify the process of recording, storing, searching and making reports. In designing a web-based inventory information system, the author uses the Rapid Application Development (RAD) method. This information system is the best solution for solving problems in managing inventory. With the use of computer data technology, managed data becomes faster, reduces time inefficiency and reduces the occurrence of errors in processing data.
The development of increasingly advanced information technology can provide many benefits for completing work quickly and accurately. One example that requires the delivery of information quickly and accurately is the field of library, this is in accordance with the function of the library which is the heart of education. Most libraries are still many who adhere to a conventional system, of course this will result in disruption of the continuity of the process of managing books in the library. Therefore, the author takes the theme of this study regarding Library Book Management Information Systems Based on Websites by using the waterfall method on software development as well as methods of observation and literature on data collection. This Information System is the best solution for problem solving in managing library books. With the use of managed computer data technology becomes faster, reducing inefficient time and reducing the occurrence of errors in processing data.
One property that is currently being glimpsed by investors is an apartment. Property consulting companies as one of the service provider companies that become a link between apartment owners and apartment enthusiasts, have an important task in terms of providing information about the assessment of the offered institutions. This study will conduct a trial on the accuracy of apartment assessment predictions using the Support Vector Machine (SVM) method, then will be compared again with the results of the accuracy of the assessment method Support Vector Machine (SVM) combined using the optimization method Particle Swarm Optimization (PSO). The results of the combination of the application of SVM and PSO are used to optimize attribute selection in apartment valuation to improve the accuracy of using the SVM method. This study shows that the Particle Swarm Optimization (PSO) Support Vector Machine (SVM) method is a pretty good method of data classification, because it can seen from the increase in accuracy of 2.84% and AUC of 0.003. Subjects (attributes) that affect apartment valuation are seen from rent prices (price range), city (apartment location), size (area), furnisihing (equipment), bedroom (number of bedrooms), bathroom (number of bathrooms) and maids badroom (number of maid rooms). The results of the attribute testing showed that city attributes (apartment locations), furnisihing (equipment) and maid badroom (number of maid rooms) greatly influenced the valuation of an apartment.
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.