MISBAH, HIBBAN RAFA (2025) PENGARUH GAYA KEPEMIMPINAN DAN DISIPLIN KERJA TERHADAP KINERJA KARYAWAN DENGAN MACHINE LEARNING. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (627kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (105kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (133kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (373kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (464kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (95kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (94kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (388kB) |
Abstract
Rapid developments in the digital era have transformed the global business landscape, including the cosmetics manufacturing sector in Indonesia. In this context, leadership effectiveness and work discipline have emerged as critical factors determining a company's competitiveness. This study comprehensively examines the influence of leadership style through machine learning. The research methodology integrates conventional quantitative approaches with advanced machine learning-based analysis techniques. Data were collected through a structured questionnaire using a Likert scale distributed to 59 employees of the production division at a leading cosmetics company, supplemented by secondary data from the company's HRIS system during the period 2022–2024. Statistical analysis was performed using multiple linear regression and path analysis, while predictive modeling implemented the Decision Tree, Random Forest, and Support Vector Machine algorithms. On average per method, in SMOTE the highest accuracy ranking was achieved by SVM (72.22%), followed by Random Forest (66.67%) and Decision Tree (61.11%). In Oversampling, SVM remained the highest (72.22%), followed by Random Forest and Decision Tree, which both recorded an accuracy of 55.56%. Meanwhile, in Undersampling, Random Forest and Decision Tree both achieved an accuracy of 55.56%, with SVM the lowest at 38.89%. The Random Forest model proved to be the most stable in producing predictions, the Decision Tree model excelled in terms of visual interpretation despite being prone to overfitting, and the SVM model was effective at narrow interclass margins. The discussion confirmed that transformational leadership style had the strongest influence on improving performance, particularly in building intrinsic employee motivation. Work discipline was shown to play a significant role as a mediator in the relationship between leadership and performance. Furthermore, the machine learning model was able to identify complex interaction patterns between variables that were not detected by traditional statistical analysis, including the indirect influence of demographic factors. The practical implications of this research include recommendations for digital competency-based leadership development programs, a real-time IoT-based work discipline monitoring system, and an HR analytics framework integrated with corporate business intelligence. Academically, this research enriches the literature on HR management 4.0 by presenting an integrative model that connects leadership theory, industrial psychology, and data science. Keywords: Leadership, Discipline, Performance, Employee, Machine Learning Perkembangan pesat di era digital telah mentransformasi lanskap bisnis secara global, termasuk sektor manufaktur kosmetik di Indonesia. Pada konteks ini, efektivitas kepemimpinan dan kedisiplinan kerja muncul sebagai faktor kritis yang menentukan daya saing perusahaan. Penelitian ini secara komprehensif mengkaji pengaruh gaya kepemimpinan melalui metode Machine Learning. Metodologi penelitian mengintegrasikan pendekatan kuantitatif konvensional dengan teknik analisis mutakhir berbasis machine learning. Data dikumpulkan melalui kuesioner terstruktur menggunakan skala Likert yang disebarkan kepada 59 karyawan divisi produksi di perusahaan kosmetik terkemuka, dilengkapi dengan data sekunder dari sistem HRIS perusahaan selama periode 2022–2024. Analisis statistik dilakukan dengan regresi linier berganda dan analisis jalur, sementara pemodelan prediktif mengimplementasikan algoritma Decision Tree, Random Forest, dan Support Vector Machine. Secara rata-rata per metode, pada SMOTE peringkat akurasi tertinggi diraih oleh SVM (72.22%), diikuti Random Forest (66.67%) dan Decision Tree (61.11%). Pada Oversampling, SVM tetap tertinggi (72.22%), diikuti Random Forest dan Decision Tree yang sama-sama mencatat akurasi 55.56%. Sementara pada Undersampling, Random Forest dan Decision Tree sama-sama meraih akurasi 55.56%, dengan SVM terendah di 38.89%. Model Random Forest terbukti paling stabil dalam menghasilkan prediksi, Decision Tree unggul dari sisi interpretasi visual meski rawan overfitting, dan SVM efektif pada margin antar kelas yang tipis. Pembahasan menegaskan bahwa gaya kepemimpinan transformasional memberikan pengaruh paling kuat terhadap peningkatan kinerja, khususnya dalam membangun motivasi intrinsik karyawan. Disiplin kerja terbukti berperan sebagai mediator penting dalam hubungan antara kepemimpinan dan kinerja. Lebih jauh, model machine learning mampu mengidentifikasi pola interaksi kompleks antar variabel yang tidak terdeteksi analisis statistik tradisional, termasuk pengaruh tidak langsung faktor demografi. Implikasi praktis penelitian ini mencakup rekomendasi untuk program pengembangan kepemimpinan berbasis kompetensi digital, sistem monitoring disiplin kerja real-time berbasis IoT, dan framework HR analytics yang terintegrasi dengan business intelligence perusahaan. Secara akademis, penelitian ini memperkaya literatur tentang manajemen SDM 4.0 dengan menyajikan model integratif yang menghubungkan teori kepemimpinan, psikologi industri, dan ilmu data. Kata Kunci: Kepemimpinan, Disiplin, Kinerja, Karyawan, Machine Learning
Actions (login required)
 |
View Item |
