#TITLE_ALTERNATIVE#

• Main Author: KURNIA SEPTIALOKA (NIM: 13514028) , DHARMA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/26575
• Institution: Institut Teknologi Bandung

Expert is a person who is recognized to have a comprehensive and authoritative knowledge of or skill in a particular area. Expert finding is important and needed both in industry and academy. One of the shortcomings of existing expert finding system is that there is no expertise validation for the expert’s skills which is selfproclaimed. Furthermore, the criteria to determine someone's expertise is portrayed poorly through their profile page and those criteria might not meet finder's needs for expertise sufficiently. Currently in Indonesia, there is no system to accommodate the need of finding expert publicly, specifically for educational institutions. <br /> <br /> <br /> Expert finding system in this research is developed by using Indonesia’s lecturers data taken by PD Dikti Indonesia website. The system consists of two main subsystem, which are classifying the expertise area based on the problems given and recommending the suitable lecture in accordance with the expertise classification result. Data is collected by crawling through the website. Classification model is built using machine learning approach with feature extraction using TF-IDF. The expert candidate’s score is calculated using weighted sum model method by which the weight and score are determined by admin. Criterias to determine expert are education, research, teaching, position, and community service (project). The system is a web application developed using Django framework. For validation, the system is tested using 3 methods, which are functionality, classification model quality, and expert recommendation quality. From the result, it can be concluded that the system requirement is satisfied. The classification model is adequate with 79% accuracy with Random Forest algorithm for 10 class classification. <br /> <br /> <br /> The system is expected to be used by government and education institution to find the expert recommendation especially lecturer for projects or community services. In the future, it is suggested that the multi-label classification model will be used so that the expertise area classification could be more accurate to address the problem given. <br />