Skripsi

PEMETAAN TIPE KEPRIBADIAN HOLLAND (RIASEC) MELALUI KLASIFIKASI MULTI-TASK FITUR GRAFOLOGI TULISAN TANGAN MENGGUNAKAN RESNEXT50

---

Graphology-based personality analysis has emerged as a growing approach in psychometrics and artificial intelligence for identifying individual characteristics through handwriting, while conventional personality assessment methods remain largely dependent on subjective self-assessment questionnaires. This study proposes a mapping of Holland’s personality types (RIASEC) through the classification of handwriting graphological features using a deep learning approach based on a ResNeXt50 Convolutional Neural Network with a multi-task learning scheme. The proposed model employs five output heads to classify size, slant, word spacing, line spacing, and pen pressure features. Data splitting is performed using Multilabel Stratified Shuffle Split with the Iterative Stratification algorithm to preserve class distribution proportions across all graphological features. The training process applies Weighted Cross-Entropy Loss to each classification head, with class weights computed based on the inverse frequency of class occurrences in the training data, and optimization is conducted using the Adam optimizer, with three independent training runs performed to ensure result reliability. Experimental results demonstrate consistent performance across all graphological features, achieving accuracies of 81.82% for size, 77.92% for slant, 83.55% for word spacing, 86.15% for line spacing, and 83.12% for pen pressure, with balanced F1-scores across features. The classified graphological features are subsequently mapped to RIASEC personality types with the output constrained to a maximum of three dominant types, indicating that the proposed multi-task ResNeXt50-based approach, combined with iterative stratification and weighted loss, offers a promising and objective alternative for personality assessment in career development and educational contexts.

---

Availability

Detail Information

Series Title

-

Call Number

T1914882025

Publisher

Indralaya : Prodi Magister Ilmu Komputer, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xv, 32 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.370 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Pengenalan pola
Prodi Magister Ilmu Komputer

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment