Website Traffic Time Series Forecasting Using Regression Machine Learning
Dhruv Sikka, C. N. S. Vinoth Kumar · 2023
The term ‘‘Web analytics’’ pertains to the act of monitoring, analysing, and creating reports regarding the use of a website, such as its web pages, images, and videos. By utilising web analytics, businesses can retain their customers, increase their website traffic, and attract more visitors, which ultimately leads to higher revenue. A study has been conducted on statistical forecasting teaching notes of a website, which involved analysing five years of daily time series data for various traffic measures. These measures included the daily counts of page views, unique visitors, first-time visitors, and returning visitors for academic teaching notes on the website. This work presents a single unit of multiple regression machine learning models that can predict new user and returning user web traffic amounts over some time that can be expected over a certain period or on a specific time interval. The voting Regression model combines algorithms like Decision Tree Regression, Multi Linear Regression and Support Vector Machine Regression into a single unit that can predict expected traffic volume over a period with 99.96% accuracy and 0.24% of absolute error.