Predicting customer behavior using indexed data from video analytics: machine learning application with stochastic gradient-bossted trees | |
| Author | Alfian, Alfian |
| Call Number | AIT RSPR no.CS-18-01 |
| Subject(s) | Video analytics Neural Networks Pattern recognition systems Machine Learning |
| Note | A research study submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Computer Science, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Research studies project report ; no. CS-18-01 |
| Abstract | Retail stores strive to maximize sales. However, maximizing sales is difficult unless they know customer behavior correlated with sales. The problem is that retail stores understand little about such behavior. Consequently, they miss the opportunity to maximize sales. To solve such a problem, patterns of customer behavior must be predicted. Therefore, this experiment conducted CCTV camera observation in a coffee shop and had video analytics index customer movements that represent customer behavior variables, such as sitting customers and coffee takeaway, for analysis. Regression analysis that retail stores usually use produces less accurate prediction results, because it cannot handle large data sets, nonlinear variables, and hourly scaled prediction. Therefore, I used stochastic gradient-boosted trees (SGBT), because studies have suggested SGBT may produce accurate results on large data sets, nonlinear variables, and hourly scaled prediction. Moreover, SGBT has never been used in retail settings before. My experiment generated prediction results only accurate for sitting customers, because data in this variable was dense. However, the prediction results for coffee takeaway were inaccurate, because data in that variable was sparse. Accordingly, SGBT is useful in businesses for predicting customer behavior variables whose data is dense, large, nonlinear, and sequential in time series |
| Year | 2018 |
| Corresponding Series Added Entry | Asian Institute of Technology. Research studies project report ; no. CS-18-01 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Dailey, Mathew N.; |
| Examination Committee(s) | Mongkol Ekpanyapong;Vatcharaporn Esichaikul; |
| Scholarship Donor(s) | Thailand (HM King); |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 2018 |