Prediction of Consumer Purchase Intention Using E-Commerce Web Data

Abstract

Digital advertising is noted as the favored technique compared to traditional advertising.
These techniques are known to be useful by people actively engaged in social media
advertising and for those who educate about social media advertising and regarding purchase
intention. This paper provides some preliminary cognizance into the consumer-views of
online advertisements on social media and buying behavior. Business, academicians,
researchers all exchange their commercials and statistics on the internet so they can
potentially be related with humans swiftly and without problems to survey on searchable
product websites by the method of web scraping. Web scraping is an approach by which a
user can extract a copious amount of web data and abundance of statistics from websites
besides the unstructured facts available on websites. This is an automated technique for the
activity of data- extraction but the problem is the unstructured nature. To prevent this
problem, internet scraping facilitates the researcher to acquire these unstructured facts and
store it in an established form. Thus, vital research vicinity is the outcome of the customer
rating and customer’s product assessment. The purpose is to review a given dataset with the
usage of machine learning based strategies for product scores and rating forecasting through
prediction results in the best accuracy. The analysis of data-set via Support Vector Machine
classifier (SVM) to capture numerous data like variable identity, univariate analysis, bivariate
analysis and multi-variate analysis, missing value treatments and probing the data
validation for the accurate data input such as type, cleaning of records and making fit for the
inspection and information visualizations will be achieved for the whole data-set. The
analysis is used for the comprehensive and detailed guide for sensitivity inspection of model
parameters towards performance in prognosis of the product rating with price details with the
aid of finding accuracy calculation. So, the proposed work is prediction of consumer
intention from a given e-commerce dataset with software of web user interface.