Penggunaan Metode Fuzzy Time Series Chen dan Fuzzy Time Series Cheng untuk Memprediksi Laba pada PT Taspen (Persero)

Fajrin, Dirani Amaris (2025) Penggunaan Metode Fuzzy Time Series Chen dan Fuzzy Time Series Cheng untuk Memprediksi Laba pada PT Taspen (Persero). S1 thesis, UNIVERSITAS JAMBI.

[img] Text
ABSTRAK.pdf
Download (188kB)
[img] Text
BAB I.pdf
Download (245kB)
[img] Text
BAB V.pdf
Download (133kB)
[img] Text
COVER.pdf
Download (625kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (156kB)
[img] Text
SKRIPSI FULL_DIRANI AMARIS FAJRIN_F1C220029.pdf
Restricted to Repository staff only
Download (1MB)
[img] Text
HALAMAN PENGESAHAN.pdf
Download (198kB)
Official URL: https://repository.unja.ac.id/

Abstract

PT Taspen (Persero), sebagai BUMN yang bergerak di bidang jasa, perlu menganalisis kinerja keuangan untuk memahami fluktuasi laba yang terjadi setiap triwulan pada tahun 2022. Laba bersih meningkat pada triwulan II karena perpanjangan usia pensiun dan kondisi pasar yang membaik, serta meningkat kembali di triwulan III karena pendapatan yang naik. Namun, triwulan IV menunjukkan penurunan laba karena tidak adanya keuntungan. Fluktuasi ini menuntut dilakukannya peramalan untuk mengetahui prospek dan prediksi dividen di masa depan. Metode peramalan yang digunakan dalam penelitian ini adalah Fuzzy Time Series (FTS), khususnya dua metode yaitu Fuzzy Time Series Chen dan Fuzzy Time Series Cheng. Fuzzy Time Series bekerja berdasarkan konsep fuzzy logic yang mengabaikan asumsi normalitas dan stasioneritas data, menjadikannya fleksibel untuk memodelkan data time series. Fuzzy Time Series Chen dikenal dengan pendekatannya yang sederhana dan mudah diterapkan, cocok untuk pengguna tanpa pemahaman mendalam tentang fuzzy logic. Fuzzy Time Series Cheng menawarkan akurasi lebih tinggi melalui penentuan interval fuzzy yang lebih kompleks, sehingga lebih efektif untuk pola data yang rumit. Tahap analisis yang dilakukan pada metode Fuzzy Time Series Chen dan Fuzzy Time Series Cheng yaitu pada Fuzzy Time Series Chen yaitu menentukan himpunan semesta (U), membentuk interval membentuk interval, mendefinisikan himpunan fuzzy, menentukan nilai keanggotaan masing-masing data, melakukan fuzzifikasi data, menentukan Fuzzy Logic Relationship (FLR), menentukan Fuzzy Logic Relationship Group (FLRG), menentukan peramalan menurut FLRG dan defuzzifikasi. Fuzzy Time Series Cheng menentukan himpunan semesta (U), membentuk interval membentuk interval, mendefinisikan himpunan fuzzy, menentukan nilai keanggotaan masing-masing data, melakukan fuzzifikasi data, menentukan Fuzzy Logic Relationship (FLR), menentukan Fuzzy Logic Relationship Group (FLRG), menentukan peramalan menurut FLRG ke dalam bentuk matriks pembobot (W), mentransfer bobot FLRG ke dalam bentuk matriks pembobot terstandarisasi (W^*) dan defuzzifikasi. Setelah didapatkan semua hasil peramalan dari kedua metode, selanjutnya melakukan ketepatan peramalan melalui tiga ukuran error, yaitu Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), dan Mean absolute Error (MAE). Berdasarkan hasil perhitungan ukuran ketepatan peramalan laba PT Taspen, diketahui bahwa metode Cheng memiliki nilai MAPE sebesar 0,603, MSE sebesar 2,684, dan MAE sebesar 1.172,53, sedangkan metode Chen menunjukkan nilai MAPE sebesar 0,674, MSE sebesar 3,303, dan MAE sebesar 1.319,29. Nilai-nilai tersebut menunjukkan bahwa metode Cheng memberikan hasil peramalan yang lebih akurat dibandingkan metode Chen, dikarenakan pada nilai MAPE, nilai MSE, dan nilai MAE menujukkan ukuran error yang lebih kecil. Secara khusus, nilai MAPE pada metode Cheng yang lebih kecil menunjukkan bahwa kesalahan relatif peramalan terhadap nilai aktual lebih kecil, sehingga metode Cheng lebih andal dan layak dipertimbangkan dalam memproyeksikan laba PT Taspen (Persero). PT Taspen (Persero), as a state-owned enterprise engaged in services, needs to analyze its financial performance to understand the profit fluctuations that occur every quarter in 2022. Net profit increased in the second quarter due to the extension of the retirement age and improved market conditions, and increased again in the third quarter due to increased revenue. However, the fourth quarter showed a decrease in profit due to the absence of profits. These fluctuations demand forecasting to determine the prospects and predictions of dividends in the future. The forecasting method used in this research is Fuzzy Time Series (FTS), specifically two methods namely Chen's Fuzzy Time Series and Cheng's Fuzzy Time Series. Fuzzy Time Series works based on the concept of fuzzy logic that ignores the assumptions of normality and stationarity of data, making it flexible to model time series data. Chen's Fuzzy Time Series is known for its simple and easy-to-implement approach, suitable for users without a deep understanding of fuzzy logic. Cheng's Fuzzy Time Series offers higher accuracy through the determination of more complex fuzzy intervals, making it more effective for complicated data patterns. The analysis stage carried out in Chen's Fuzzy Time Series and Cheng's Fuzzy Time Series methods, namely in Chen's Fuzzy Time Series, namely determining the universe set (U) forming intervals, defining fuzzy sets, determining the membership value of each data, fuzzifying data, determining the Fuzzy Logic Relationship (FLR), determining the Fuzzy Logic Relationship Group (FLRG), determining forecasting according to FLRG and defuzzification. Cheng's Fuzzy Time Series determines the universe set (U), forms intervals, defines fuzzy sets, determines the membership value of each data, fuzzifies data, determines Fuzzy Logic Relationship (FLR), determines Fuzzy Logic Relationship Group (FLRG), determines forecasting according to FLRG into the form of a weight matrix (W), transfers FLRG weights into the form of a standardized weight matrix (W^* ) and defuzzification. After obtaining all the forecasting results from the two methods, then perform forecasting accuracy through three error measures, namely Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), and Mean absolute Error (MAE). Based on the calculation of the accuracy of PT Taspen's profit forecasting, it is known that Cheng's method has a MAPE value of 0.603, MSE of 2.684, and MAE of 1,172.53, while Chen's method shows a MAPE value of 0.674, MSE of 3.303, and MAE of 1,319.29. These values show that Cheng's method provides more accurate forecasting results than Chen's method, because the MAPE value, MSE value, and MAE value show smaller error sizes. In particular, the smaller MAPE value in Cheng's method indicates that the relative error of the forecast to the actual value is smaller, so Cheng's method is more reliable and worth considering in projecting the profit of PT Taspen (Persero).

Type: Thesis (S1)
Uncontrolled Keywords: Fuzzy Time Series Chen, Fuzzy Time Series Cheng, Prediksi Laba
Subjects: L Education > L Education (General)
Divisions: Fakultas Sains dan Teknologi > Matematika
Depositing User: FAJRIN
Date Deposited: 14 Oct 2025 06:37
Last Modified: 14 Oct 2025 06:37
URI: https://repository.unja.ac.id/id/eprint/86596

Actions (login required)

View Item View Item
Repository Unja is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.