Abstract
In application development, achieving synergy between evolving software and elicitation process outcomes emerges as a foundational objective. This endeavor is empowered by the role of text analysis, a tool that adeptly uncovers intricate data relationships and contextual intricacies. This is especially pertinent given the textual attributes of System Requirements Specification (SRS) information. Optical Character Recognition (OCR) adds depth to comprehension by converting images into text. However, challenges surface due to variances between functional requirements and the design's control and affordance components, arising from divergent interpretations of functional requirements during elicitation. Our focal point resides in crafting an application to evaluate the alignment between UI design and the textual core of functional requirements. This study reveals 13 Functional Requirements (d1-d13) and categorizes two control/affordance component documents (d14-d15). The integration of text analysis, steered by word2vec, drives an alignment process, culminating in a notable 0.397 score. Recommendations pivot on the Cohen's Kappa index; values below 0.21 prompt adjustments for alignment at a "sufficient" level (0.21 or higher). The culmination of research materializes in an advanced application adept at processing txt files, skillfully extracting inputs via text analysis, and semantically evaluating alignment through Word2Vec. It offers insights for data instances where standards fall short, firmly rooted in Cohen's Kappa index principles. This application functions as a comprehensive repository, methodically outlining enhancement suggestions, assessed against the backdrop of Cohen's Kappa index.
Keywords — Control and affordance components, User Interface, Text analysis, Functional Requirements, Semantic Similarities
Abstrak
Dalam pengembangan aplikasi, artefak harus saling berkaitan agar sesuai dengan proses elisitasi. Kesesuaian dapat dicari dengan metode text analysis, mengekstrak dan menganalisis data untuk mewakili isi konten teks SRS. Metode Optical Character Recognition (OCR) digunakan untuk mendapatkan informasi teks pada komponen desain UI. Namun, terdapat masalah ketidaksesuaian antara functional requirement dengan komponen control and affordances pada UI yang telah terdefinisi karena perbedaan saat menentukan functional requirement dari hasil elisitasi. Penelitian ini bertujuan mengembangkan aplikasi yang mengukur kesesuaian informasi teks dari desain UI dengan functional requirement. Hasil penelitian berupa 13 Functional requirement (d1-d13) dan 2 komponen control and affordance (d14-d15). Melalui text analysis, dilakukan kesesuaian menggunakan word2vec dengan skor tertinggi 0,397 . Rekomendasi perbaikan dilakukan berdasarkan indeks cohen kappa, nilai kesesuaian di bawah 0,21 harus diperbaiki agar kedua dataset memiliki kesesuaian "cukup" (?0,21). Penelitian ini sudah menghasilkan aplikasi yang dapat menginput dan membaca file txt, mengekstrak input dengan text analysis, dan mengukur kesesuaian secara semantik menggunakan Word2Vec. Aplikasi ini akan memberikan tabulasi untuk rekomendasi perbaikan dari data dengan nilai di bawah standar berdasarkan indeks cohen kappa.
Kata kunci: Komponen Control Affordance, User Interface, Text analysis, Functional Requirements, Similaritas Semantic