The main purpose of this research is to create a predictive model of the ARIMA method on the daily stock price of PT. Garuda Indonesia, Tbk during the Covid-19 pandemic. This study uses daily secondary data from April 22, 2019, to April 20, 2020. The results of research using the ARIMA model shows that data from April 22, 2019, to April 20, 2020, can be used to predict stock closing prices from April 21, 2020, to July 13, 2020. The ARIMA model obtained the results of daily stock price predictions of PT. Garuda Indonesia, Tbk on the Indonesia Stock Exchange from 21 April 2020 to 13 July 2020 tend to experience a decline. This is presumably because investors tend to hold back their capital due to the government's prohibition on going home, which resulted in the cessation of operations in the aviation sector. Keywords: Covid-19, Garuda Indonesia, stock price, ARIMA method.

The purpose of this study is to classify the student's length of study based on the status of graduating on time or not on time based on several independent variables observed, namely gender, Grade Point Average (GPA), place of residence, type of parents occupation and school origin. The statistics used in this study is non-parametric statistics with a classification analysis method. The classification analysis is to find a training set model of the training set that distinguishes records into appropriate categories or classes. The method used is classification using ensemble techniques. The basic principle of the ensemble method is to develop a set of models from training data and combine a set of models to determine the final classification. The final classification is based on the largest collection of votes from a combination of a set of models. To get the best combination of models, the ensemble method enables the use of several different classification models. The ensemble method used in this study is Bagging and Boosting. Keywords: Ensemble Analysis, Classification, Bagging, Boosting, Students Length of Study, Indonesia.

Uang elektronik tumbuh secara pesat di Indonesia. Termasuk beberapa di antaranya yang menempati urutan teratas paling banyak digunakan oleh masyarakat adalah Gopay, OVO, dan ShopeePay. Adapun penelitian ini ditujukan guna melakukan identifikasi adanya efek persepsi kemudahan penggunaan pada minat penggunaan uang elektronik (Gopay, OVO, dan ShopeePay). Penelitian ini berjenis penelitian kausalitas dengan pendekatannya adalah pendekatan kuantitatif. Populasi yang ditetapkan merupakan mahasiswa jurusan Manajemen dan jurusan Akuntansi Universitas Klabat. Dan melalui purposive sampling, ditetapkan adanya persyaratan responden bahwa responden harus memiliki akun Gopay, OVO, atau ShopeePay. Kemudian didapati sebanyak 58 sampel yang memenuhi ketentuan dan telah mengisi kuesioner yang disebar dengan sistem online. Data yang telah dikumpulkan, selanjutnya dianalisis dengan analisis regresi sederhana. Temuan analisis membuktikan adanya efek persepsi kemudahan penggunaan secara signifikan dan positif pada minat penggunaan uang elektronik (Gopay, OVO, dan ShopeePay) pada mahasiswa jurusan Manajemen dan jurusan Akuntansi Universitas Klabat.

The purpose of this study is to classify the student’s length of study based on the status of graduating on time or not on time based on several independent variables observed, namely gender, Grade Point Average (GPA), place of residence, type of parents occupation and school origin. The statistics used in this study is non-parametric statistics with a classification analysis method. The classification analysis is to find a training set model of the training set that distinguishes records into appropriate categories or classes. The method used is classification using ensemble techniques. The basic principle of the ensemble method is to develop a set of models from training data and combine a set of models to determine the final classification. The final classification is based on the largest collection of votes from a combination of a set of models. To get the best combination of models, the ensemble method enables the use of several different classification models. The ensemble method used in this study is Bagging and Boosting.