Article Preview
Top1. Introduction
For the survival of a business, stability and growth are very important aspects. To survive is one of the most important things for a business because of frequent and unexpected turbulence in the market. Particularly for a large business handling global competitions, awareness of users and ever-increasing consumers’ demands knowledge of consumer behaviour has become critical. Knowledge of consumer behaviour can help in identifying new marketing opportunities, possibilities to gain competitive advantages, help consumers in getting the product and services they desire, and add value to the business operations. In the era of information and communication technology, lots of data are generated during operation through inventory, production, marketing, sales, finance, and other related functions. These data must be utilised for the betterment of the business. It has been observed that data for transactions related to a business come from varieties of locations, devices, and formats. Further, a business generally has multiple products and services. Hence, there are a significant amount of data about consumers, product, services, policies, banking and insurance details, and other related activities for a business. These data are offline as well as online. Such data need to be effectively utilised to help managers and business professionals to identify consumer needs, understand consumer behaviour, and strengthen the business-consumer relationship. For a business, success is more than profit; its inclination is more towards the satisfaction of the consumers, which can be enhanced by improving offerings to the consumers. By efficiently managing data related to the domain, professionals and managers can improve their practices and refine their delivery of products and services with the right product and right message to the right consumer at the right time. It will be unfortunate if such big consumer data are not further analysed to improve a business.
Issues raised related to knowledge discovery, modelling, and handling voluminous data on various web platforms can be effectively handled by intelligent techniques compared to traditional data-based systems (Akerkar & Sajja, 2016). Artificial Intelligence and Machine Learning (AI and ML) methods are useful for handling a large amount of data related to consumers, products, and transactions and discovering knowledge about consumer behaviour patterns. Such consumer behaviour patterns do not follow a generic formula and logic but often follow evolutionary and natural approaches. Further, with the use of fuzzy logic, uncertainty related to the domain can also be handled effectively, as suggested by Priti Sajja (2020). The proposed work highlights the application of the hybrid genetic fuzzy system for effective consumer modelling and knowledge discovery. A generic architecture is proposed with multiple phases. The proposed work can be used in various domains where there is a need to model consumer behaviour and discover insights about consumers. Knowledge of consumers’ behaviour leads to the success of the business and strengthens the business, and often leads to other businesses. Besides tracking the possible consumers, one can identify the possible categories of the consumers and future products, improve the quality of products and services, and improve the brand image of the company.
The architecture described in this work proposes the use of a genetic algorithm to encode fuzzy rules using the binary encoding strategy and demonstrates the evolution of these rules automatically. The article also illustrates how fitness can be calculated, and the consumer behaviour pattern is stored in the form of encoded fuzzy rules. Instead of manually deducing and generalising the consumer behaviour pattern, strong, valid, and necessary fuzzy rules are automatically evolved. This article describes work done so far in the area with common observations, presents a generic architecture, example rules, fuzzy membership functions, and evolution of strong rules with fitness function. The improvement between generations is also shown using the graph showing the fitness of the populations. In the end, the article is concluded with applications and possible future enhancements to work.