Hello, I am applying for the 2019 Korean Government Scholarship Program (KGSP) for Graduate Studies and this is my Goal of Study & Study Plan. I hope someone would take the time to help me fix a few grammatical errors, and if there's a need to edit the content and structure, or something I could add to strengthen my case.
Goal of study, title or subject of research, and detailed study plan
The reason I have chosen to study a master's degree in industrial and management engineering is to progress my career towards the industrial application of machine learning. I am hoping to further strengthen my foundation in mathematics and research skills relevant to data science. At the end of the program, I envision myself to be more capable of promoting a data-driven culture in my local community by combining engineering knowledge with business administration in order to improve productivity. In the long run, I want to be able to build a solid reputation for consultancy and help organizations achieve optimization and continuous improvement through the effective utilization of data.
Data mining is the subject of research I am interested in and has long fascinated me since my senior year in the university. I developed a keen interest in data mining particularly on clustering algorithms when I was working on my undergraduate thesis. Clustering is an unsupervised machine learning algorithm that has the goal of decomposing a given set of objects based on similarity or dissimilarity - to model the underlying structure or data distribution in the data in order to learn more about the data. There are numerous ways of categorizing clustering algorithms - partitional, hierarchical, distance-based, or density-based - according to result or according to optimization. According to optimization, a density-based clustering algorithm can be mathematically modeled as a multimodal optimization problem, which can come off as NP-hard optimization problem.
Metaheuristic algorithms are effective in addressing the computational complexity brought about by NP-hard optimization problems because they find a near-optimal solution in the search space that takes up the least amount of time. The two main features of metaheuristic algorithms are exploration and exploitation. Exploration enables the algorithm to escape local optimum traps in finding solutions in the search space while exploitation carefully examines a promising region within the search space to find the best optimum. A metaheuristic algorithm is said to be effective if there is a balance between these two features.
Clustering and metaheuristics are significant in machine learning. Since algorithms need to be trained to learn how to classify and process information, the efficiency and accuracy of the algorithm depend on how well the algorithm was trained. Training an algorithm takes time and CPU resources - the more so if the dataset is large and multidimensional. Thus, research on clustering techniques that are scalable on large multidimensional data is important in order to reduce the computation time and utilize far fewer CPU resources. I want to take part in research on data mining, specifically on developing new clustering algorithms that are scalable on the large multidimensional dataset using metaheuristics.
Since the MSIE program requires an overall 28 credits, four of which is compulsory for research, I would like to divide it to the following courses: three basic courses (500 level courses), four advanced courses (600 level courses), an independent study in a specific research area (700 level courses), and a research course. My end goal is to take Master Thesis Research (IMEN699) but I will be needing a stronger foundation in statistics and intelligent systems during my first year to prepare for my research. With that, I choose the following subjects: Design and Analysis of Experiments (IMEN 542), Time Series Analysis (IMEN677), Expert Systems (IMEN584), and Advanced Artificial Intelligence (IMEN683). Also, I would like to take Information Modeling (IMEN695) and Distributed Information System (IMEN781) to gain a thorough understanding of the data appropriate to the needs of today's data science. Lastly, I want to take the opportunity of studying at POSTECH to learn about entrepreneurial related courses in industrial and management engineering such as Product Development Strategy (IMEN595) and Technology Planning (IMEN611) to learn about Korean case studies on company leaders making decisions and their outcomes, as well as their strategies in technology and innovation.
Goal of study & Study Plan
Goal of study, title or subject of research, and detailed study plan
The reason I have chosen to study a master's degree in industrial and management engineering is to progress my career towards the industrial application of machine learning. I am hoping to further strengthen my foundation in mathematics and research skills relevant to data science. At the end of the program, I envision myself to be more capable of promoting a data-driven culture in my local community by combining engineering knowledge with business administration in order to improve productivity. In the long run, I want to be able to build a solid reputation for consultancy and help organizations achieve optimization and continuous improvement through the effective utilization of data.
Data mining is the subject of research I am interested in and has long fascinated me since my senior year in the university. I developed a keen interest in data mining particularly on clustering algorithms when I was working on my undergraduate thesis. Clustering is an unsupervised machine learning algorithm that has the goal of decomposing a given set of objects based on similarity or dissimilarity - to model the underlying structure or data distribution in the data in order to learn more about the data. There are numerous ways of categorizing clustering algorithms - partitional, hierarchical, distance-based, or density-based - according to result or according to optimization. According to optimization, a density-based clustering algorithm can be mathematically modeled as a multimodal optimization problem, which can come off as NP-hard optimization problem.
Metaheuristic algorithms are effective in addressing the computational complexity brought about by NP-hard optimization problems because they find a near-optimal solution in the search space that takes up the least amount of time. The two main features of metaheuristic algorithms are exploration and exploitation. Exploration enables the algorithm to escape local optimum traps in finding solutions in the search space while exploitation carefully examines a promising region within the search space to find the best optimum. A metaheuristic algorithm is said to be effective if there is a balance between these two features.
Clustering and metaheuristics are significant in machine learning. Since algorithms need to be trained to learn how to classify and process information, the efficiency and accuracy of the algorithm depend on how well the algorithm was trained. Training an algorithm takes time and CPU resources - the more so if the dataset is large and multidimensional. Thus, research on clustering techniques that are scalable on large multidimensional data is important in order to reduce the computation time and utilize far fewer CPU resources. I want to take part in research on data mining, specifically on developing new clustering algorithms that are scalable on the large multidimensional dataset using metaheuristics.
Since the MSIE program requires an overall 28 credits, four of which is compulsory for research, I would like to divide it to the following courses: three basic courses (500 level courses), four advanced courses (600 level courses), an independent study in a specific research area (700 level courses), and a research course. My end goal is to take Master Thesis Research (IMEN699) but I will be needing a stronger foundation in statistics and intelligent systems during my first year to prepare for my research. With that, I choose the following subjects: Design and Analysis of Experiments (IMEN 542), Time Series Analysis (IMEN677), Expert Systems (IMEN584), and Advanced Artificial Intelligence (IMEN683). Also, I would like to take Information Modeling (IMEN695) and Distributed Information System (IMEN781) to gain a thorough understanding of the data appropriate to the needs of today's data science. Lastly, I want to take the opportunity of studying at POSTECH to learn about entrepreneurial related courses in industrial and management engineering such as Product Development Strategy (IMEN595) and Technology Planning (IMEN611) to learn about Korean case studies on company leaders making decisions and their outcomes, as well as their strategies in technology and innovation.
