Garuda - Garba Rujukan Digital

CROMOSOM: APLIKASI CROWDSOURCED SOFTWARE ENGINEERING MENGGUNAKAN KOMPONEN REKOMENDASI TASK BERBASIS MEDIA SOSIAL Putra, Kurnia Ramadhan; Catur Candra, Muhammad Zuhri
Rekayasa Hijau : Jurnal Teknologi Ramah Lingkungan Vol 3, No 3 (2019)
Publisher : Lembaga Penelitian dan Pengabdian Masyarakat Institut Teknologi Nasional

ABSTRAK Konsep crowdsourcing dapat dimanfaatkan pada bidang rekayasa perangkat lunak yang dikenal dengan Crowdsourced Software Engineering (CSE). CSE digunakan untuk menyelesaikan task yang berkaitan dengan perangkat lunak, seperti desain, implementasi, dan pengujian perangkat luna serta perbaikan bug. Saat ini, permasalahan umum yang terjadi aplikasi CSE adalah worker menghabiskan waktu untuk menemukan task yang relevan sesuai dengan keahliannya dan requester sulit untuk memilih worker yang dapat dipercaya untuk mengerjakan task. Komponen sistem rekomendasi dapat diintegrasikan pada aplikasi CSE yang dipercaya mampu untuk mengatasi permasalahan tersebut. Beberapa penelitian yang ada tentang integrasi komponen sistem rekomendasi pada aplikasi CSE hanya untuk menangani rekomendasi task dan tidak memepertimbangkan trustworthiness dari worker yang akan mengerjakan task. Sedangkan pada penelitian ini, integrasi komponen sistem rekomendasi selain dapat membantu merekomendasikan task kepada worker juga melakukan perangkingan terhadap worker berdasarkan trustworthiness dari worker tersebut sehingga dapat menjadi pertimbangan untuk requester dalam memilih worker yang akan mengerjakan task. Pendekatan yang diusulkan pada penelitian ini adalah kombinasi antara pendekatan content based dengan individual based. Pendekatan content based untuk menangani proses pencocokan antara kebutuhan keahlian yang diperlukan untuk mengerjakan task dengan kualifikasi worker yang akan mengerjakan task. Sedangkan pendekatan individual based untuk menangani proses perhitungan nilai social profile dalam menghasilkan trustworthiness dari worker. Implementasi dilakukan dengan mengembangkan aplikasi CSE yang dikenal dengan Cromosom, yang diintegrasikan dengan komponen sistem rekomendasi untuk membantu merekomendasikan task kepada worker dan melakukan perangkingan terhadap worker berdasarkan trustworthiness dari worker tersebut. Dari hasil pengujian fungsionalitas yang dilakukan, aplikasi Cromosom dapat membantu worker untuk menemukan task yang lebih relevan sesuai dengan keahliannya, dan membantu requester dalam memilih worker yang memiliki trustworthiness untuk mengerjakan task. Kata kunci: crowdsourcing, crowdsourced software engineering, task recommendation. ABSTRACT The concept of crowdsourcing can be utilized in the field of software engineering known as Crowdsourced Software Engineering (CSE). CSE is used to address tasks related to software, such as design, implementation, and testing of software and bug fixes. Currently, a common problem that occurs with CSE applications is that workers spend time finding relevant tasks according to their expertise and requester is difficult to choose workers who can be trusted to do the task. The recommendation system component can be integrated into CSE applications that are believed to be able to overcome these problems. Some existing research on the integration of recommendation system components in CSE applications is only to handle task recommendations and not consider the trustworthiness of workers who will be working on tasks. Whereas in this study, the integration of recommendation system components in addition to being able to help recommend tasks to workers also rank workers based on the trustworthiness of the workers so that they can be considered for the requester in choosing workers who will do the task. The approach proposed in this study is a combination of content-based and individual-based approaches. Content-based approach to handle the matching process between the skill requirements needed to do the task and the qualifications of the worker who will be working on the task. While the individual-based approach to handle the process of calculating the value of social profiles in generating trustworthiness from workers. Implementation is done by developing a CSE application known as a Cromosom, which is integrated with the recommendation system component to help recommend tasks to workers and rank workers based on the trustworthiness of the worker. From the results of the functionality testing, the Cromosom application can help workers find more relevant tasks according to their expertise, and help requester in choosing workers who have trustworthiness to do the task. Keywords: crowdsourcing, crowdsourced software engineering, task recommendation.

Pemanfaatan Metode Collaborative Filtering dengan Algoritma KNN pada Sistem Rekomendasi Produk PUTRA, KURNIA RAMADHAN; RAHMAN, ILHAM FATHUR
MIND (Multimedia Artificial Intelligent Networking Database) Journal Vol 9, No 1 (2024): MIND Journal
Publisher : Institut Teknologi Nasional, Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26760/mindjournal.v9i1.113-123

AbstrakSalah satu permasalahan customer pada e-commerce adalah sulitnya menemukan produk yang diinginkan untuk dibeli. Sistem rekomendasi mampu menangani permasalahan tersebut dengan cara mengalisis data profil customer untuk menyaring produk yang sesuai dengan profil customer kemudian merekomendasikannya kepada customer tersebut. Untuk mengetahui hubungan antara produk dengan pengguna maka dapat memanfaatkan sistem rekomendasi. Ada beberapa permasalahan pada sistem rekomendasi yaitu sparsity data, missing value, dan duplikasi data yang sering ditemukan pada data berbasis rating seperti pada e-commerce. Untuk menyelesaikan masalah ini, maka diusulkan metode Item-based Collaborative Filtering dan algoritma K-Nearest Neighbor (KNN) dengan hasil evaluasi nilai MAE sebesar 1,05 dan RMSE sebesar 1,36 yang mampu menangani sistem rekomendasi dengan baik dengan tingkat kesalahan yang kecil.Kata kunci: recommendation system, item-based collaborative filtering, KNN, Sparsity Data, Cold-Start.AbstractIn e-commerce, one common customer problem is difficulty in finding the product they want to buy. This issue can be addressed through a recommendation system, which analyzes customer profile data to filter products that match the customer's profile and then recommends them. One way to establish the relationship between products and users is by using a recommendation system. However, recommendation systems often encounter problems such as data sparsity, missing values, and data duplication, particularly in rating-based data. To address these issues, the Item-based Collaborative Filtering method and the K-Nearest Neighbor (KNN) algorithm are proposed. Evaluation results show that these methods have MAE values of 1.05 and RMSE of 1.36, indicating their effectiveness in handling the recommendation system with a low error rate.Keywords: recommendation system, collaborative filtering, item-based CF, KNN

Influence of Data Scaling and Train/Test Split Ratios on LightGBM Efficacy for Obesity Rate Prediction FAHRUDIN, NUR FITRIANTI; PUTRA, KURNIA RAMADHAN; UMAROH, SOFIA; LAUTAN, GAMAS BLOORY
MIND (Multimedia Artificial Intelligent Networking Database) Journal Vol 9, No 2 (2024): MIND Journal
Publisher : Institut Teknologi Nasional, Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26760/mindjournal.v9i2.220-234

AbstrakNormalisasi adalah proses yang tidak dapat dilewatkan dalam data mining yang membantu menyesuaikan nilai atribut data ke skala yang sama. Dalam konteks data mining, perbedaan skala antar atribut dapat menyebabkan kesalahan dalam pemodelan atau interpretasi hasil. Penggunaan normalisasi dalam pra-pemrosesan masih diperdebatkan, terutama ketika menggunakan algoritma dari kelompok pohon keputusan. Penelitian ini membandingkan model dengan data yang dinormalisasi dan tidak dinormalisasi dengan menggunakan metode normalisasi, MinMaxScaler, MaxAbsScaler, dan RobustScaler. Hasil penelitian menunjukkan bahwa model LightGBM tanpa normalisasi memiliki tingkat akurasi sebesar 96,6 dalam mengklasifikasikan tingkat obesitas pada data saat ini. Tidak hanya normalisasi yang mempengaruhi hasil klasifikasi, tetapi juga jumlah rasio antara data pelatihan dan pengujian. Penelitian menunjukkan bahwa semakin besar persentase data yang digunakan untuk pelatihan, semakin tinggi tingkat akurasinya. Pada dataset obesitas, rasio 80:20 memiliki akurasi hingga 97%.Kata kunci: Decision Tree, LightGBM, Obesitas, Data Mining, KlasifikasiAbstractNormalization is an essential process in data mining that helps adjust the values of data attributes to the same scale. In data mining, differences in attribute scales can lead to errors in modeling or interpreting results. Normalization in preprocessing is still debated, particularly when using algorithms from the decision tree family. This study compares models with normalized and non-normalized data using normalization methods such as MinMaxScaler, MaxAbsScaler, and RobustScaler. The results show that the LightGBM model without normalization achieved an accuracy rate of 96.6% in classifying obesity levels in the current dataset. Not only does normalization affect classification results, but the ratio between training and testing data also plays a role. The study indicates that the larger the percentage of data used for training, the higher the accuracy rate. In the obesity dataset, an 80:20 ratio resulted in an accuracy rate of up to 97%.Keywords: Decision Tree, LightGBM, Obesity, Data Mining, Classification

Perbandingan Metode Content-based, Collaborative dan Hybrid Filtering pada Sistem Rekomendasi Lagu PUTRA, KURNIA RAMADHAN; RACHMAN, MOHAMMAD ADITIYA
MIND (Multimedia Artificial Intelligent Networking Database) Journal Vol 9, No 2 (2024): MIND Journal
Publisher : Institut Teknologi Nasional, Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26760/mindjournal.v9i2.179-193

AbstrakSistem rekomendasi dapat dimanfaatkan untuk membantu pengguna menemukan item atau informasi sesuai preferensi mereka, termasuk lagu. Metode seperti Collaborative Filtering (CF), Content-Based Filtering (CBF), dan Hybrid Filtering digunakan untuk meningkatkan kualitas rekomendasi berdasarkan interaksi pengguna dan karakteristik konten. Penelitian ini membandingkan efektivitas ketiga metode tersebut dalam rekomendasi lagu menggunakan dataset dengan 68.330 entri data. Metode CF dan CBF diterapkan secara terpisah, lalu dikombinasikan dalam pendekatan hybrid untuk mengevaluasi peningkatan hasil. CF mencapai presisi 49.9%, CBF 39.5%, sedangkan hybrid CF-CBF mencatat presisi tertinggi sebesar 50.7%. Sebaliknya, hybrid CBF-CF menghasilkan presisi terendah, yaitu 38.4%. Kesimpulannya, pendekatan hybrid CF-CBF lebih unggul dalam merekomendasikan lagu sesuai preferensi pengguna dibandingkan metode lainnya.Kata kunci: sistem rekomendasi, rekomendasi lagu, content-based filtering, collaborative filtering, hybrid filtering AbstractRecommender systems can be utilized to assist users in discovering items or information that align with their preferences, including music. Methods such as Collaborative Filtering (CF), Content-Based Filtering (CBF), and Hybrid Filtering enhance recommendation quality based on user interactions and content characteristics. This study compares the effectiveness of these three methods in music recommendation using a dataset containing 68,330 entries. CF and CBF were implemented separately and combined in a hybrid approach to evaluate performance improvements. CF achieved a precision of 49.9% and CBF 39.5%, while the hybrid CF-CBF approach recorded the highest precision at 50.7%. In contrast, the hybrid CBF-CF approach yielded the lowest precision, at 38.4%. In conclusion, the hybrid CF-CBF approach outperforms other methods in delivering music recommendations tailored to user preferences.Keywords: recommendation system, song recommendation, content-based filtering, collaborative filtering, hybrid filtering

Implementasi Metode Analytical Hierarchy Process untuk Mendukung Pengambilan Keputusan Perekrutan Calon Dosen Putra, Kurnia Ramadhan; Reza, Flandi Muhamad
Rekayasa Hijau : Jurnal Teknologi Ramah Lingkungan Vol 9, No 2 (2025)
Publisher : Institut Teknologi Nasional, Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26760/jrh.v9i2.134-149

AbstrakPerekrutan dosen yang berkualitas merupakan langkah strategis dalam meningkatkan mutu pendidikan di perguruan tinggi. Namun, proses seleksi ini sering kali kompleks karena melibatkan berbagai kriteria yang harus dipertimbangkan secara objektif. Kesalahan dalam pengambilan keputusan dapat berdampak pada kualitas tenaga pengajar yang direkrut. Penelitian ini mengimplementasikan metode Analytical Hierarchy Process (AHP) untuk mendukung pengambilan keputusan dalam perekrutan calon dosen di Institut Teknologi Nasional (Itenas). Metode AHP digunakan untuk menganalisis bobot prioritas dari berbagai kriteria seleksi, seperti kompetensi akademik, pengalaman mengajar, publikasi ilmiah, dan kemampuan komunikasi. Hasil penelitian menunjukkan bahwa metode AHP mampu memberikan peringkat calon dosen berdasarkan bobot prioritas yang telah ditentukan dengan tingkat konsistensi rasio (CR) sebesar 0,077 yang berada dalam batas yang dapat diterima (<0,10). Implementasi metode ini membantu meningkatkan transparansi dan objektivitas dalam proses seleksi, sehingga mendukung institusi dalam memilih kandidat yang paling sesuai dengan kebutuhan akademik dan visi institusi.Kata kunci: perekrutan dosen, sistem pendukung keputusan, analytical hierarchy process, consistency index, consistency ratio AbstractThe recruitment of high-quality lecturers is a strategic step in enhancing the quality of education in higher education institutions. However, the selection process is often complex as it involves various criteria that must be considered objectively. Errors in decision-making can impact the quality of the recruited teaching staff. This study implements the Analytical Hierarchy Process (AHP) method to support decision-making in the recruitment of prospective lecturers at the National Institute of Technology (Itenas). The AHP method is used to analyze the priority weights of various selection criteria, such as academic competence, teaching experience, scientific publications, and communication skills. The results indicate that the AHP method can rank prospective lecturers based on the predetermined priority weights, with a consistency ratio (CR) of 0.077, which is within the acceptable threshold (<0.10). The implementation of this method helps enhance transparency and objectivity in the selection process, thereby supporting the institution in selecting the most suitable candidates for academic needs and institutional vision.Keywords: lecturer recruitment, decision support system, analytical hierarchy process, consistency index, consistency ratio

Comparison of Prediction Models: Decision Tree, Random Forest, and Support Vector Regression Putra, Kurnia Ramadhan
Jurnal Informatika dan Rekayasa Perangkat Lunak Vol. 6 No. 1 (2025): Volume 6 Number 1 March 2025
Publisher : Universitas Teknokrat Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33365/jatika.v6i1.18

The Information Technology (IT) industry continues to grow rapidly, creating challenges in determining fair and competitive salaries for professionals. Accurate salary predictions are essential for companies to attract and retain talent while providing insights for individual career planning. This research aims to compare the performance of three machine learning models, such as Decision Tree Regression, Random Forest Regression, and Support Vector Regression in predicting IT sector salaries using demographic and professional data, including age, gender, education level, job position, and work experience. The study uses a dataset of 6,704 entries from Kaggle, with relationships between variables analyzed through statistical techniques such as Pearson Correlation and ANOVA. Model performance was evaluated using the R² Score, Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). Among the models, Random Forest Regression demonstrated the best performance, achieving the highest R² of 91.49% and an RMSE of 0.058, indicating high predictive accuracy with low error rates. Scatter plot visualizations confirm a strong correlation between actual and predicted salaries, supported by error analysis identifying minimal overestimation and underestimation cases. The research concludes that Random Forest Regression is the most effective model for IT salary predictions. These findings provide practical insights for organizations and individuals, highlighting the potential of data-driven approaches in salary determination. Future studies may focus on hyperparameter optimization and incorporating additional features to improve model performance and generalizability further improve model performance and generalizability.

Title

Found 6 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 6 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 6 Documents
Search