Expert Systems for Marketing

Introduction

Since the invention of computers and machines, their performance capabilities have increased exponentially. This evaluation recognized the emergence of a branch of computing called artificial intelligence, which pursues the creation of computers or machines as intelligent as human beings. AI is evolving at such an astonishing speed that it sometimes seems magical. Humans have developed AI systems by introducing all kinds of intelligences to reduce the effort and time required for maintenance personnel to carry out tedious tasks.

Artificial intelligence concepts and tools are already present in many areas, including business. Companies have shown particular interest in "expert systems," which can help them make decisions related to finance, production, marketing, and accounting. These expert systems have generated considerable enthusiasm in the professional world because they allowed the computerization of certain functions such as identification, situation diagnosis, event planning, object design, and action planning.

These applications are specifically manifested in businesses, showing that the expert system used as a tool to help them also significantly improved the effectiveness and efficiency of marketing decision-making. In order to identify the usefulness of expert systems, it is necessary to first identify their concept, components, types and methods of construction, mechanism of functioning, differences from traditional systems, and the difficulties of their use.

1. Expert Systems

The development of expert systems represents an important part of the field of artificial intelligence (AI), which aims to simulate, through machines, the reasoning capacities of human beings (e.g., voice recognition).

1.1 Definition

An expert system is a system that uses knowledge in a complex and specific field of application and acts as an expert advisor to users (Blili, 1997). It can also be defined as an information system exploiting explicit knowledge relating to a particular domain (O'Brien, 1999). Furthermore, an expert system is a computer system that mimics the decision-making ability of a human expert. It allows you to model reasoning, manipulate knowledge in a declarative form, facilitate its acquisition, modification, and updating, obtain deductions and conclusions, and produce explanations on how the results are obtained.

How It Works

The user's question is read by the expert system and translated by the natural language translation interface.
The meaning of the question is interpreted by the inference engine using the knowledge base.
A strategy is implemented to arrive at one or more solutions.

Programming Expert Systems

Expert systems can be programmed in different ways:

Using standard (procedural) programming languages, such as PASCAL or C.
Using AI-oriented programming languages like PROLOG or LISP.
Using an existing expert system infrastructure (shells) such as EMYCIN, STAGE, or EXPERTEASE.
Using languages specialized in knowledge representation, such as ROSIE or KRL.

1.2 History

Expert systems are one of the first and most important applications of artificial intelligence. Since the development of the first knowledge and expertise-based system for successful chemical analysis, the idea of producing systems based on human experience began to spread rapidly to other fields.

Notable Developments

The first DENDRAL expert system in spectrographic analysis was developed by Fiegenbaum at the end of the 1960s. This program was designed to simulate the work of a chemist reconstituting the structural formula of an organic component from its raw formula and the results of its mass spectrography.
In the field of medical diagnosis, the MYCIN system was developed in 1976 to diagnose certain infectious diseases.
In geological prospecting, the PROSPECTOR program was based on the concepts defined by Fiegenbaum in DENDRAL.

The use of expert systems has experienced accelerated growth since the 1980s, as reflected by the increase in the number of conferences on the subject, the development of programming languages and tools for creating expert systems, and the growing mastery of these techniques.

1.3 Characteristics, Capabilities, and Advantages

Expert systems are useful in a wide range of situations:

Need to diagnose problems in a given context.
Need to understand the nature of a particular situation.
Need to predict the repercussions of a current or future event.
Need to control a particular process or activity.
Need recommendations or a solution for a particular problem.

Important Characteristics of AI Expert Systems

Highest Level of Expertise: The AI expert system offers the highest level of expertise, providing efficiency, precision, and imaginative problem-solving.
Just-in-Time Reaction: An AI expert system interacts within a reasonable deadline, providing solutions faster than a human expert.
Good Reliability: The AI expert system must be reliable and free from errors.
Flexibility: It is essential that the expert system remains flexible.
Effective Mechanism: The AI expert system must have an effective mechanism for managing the compilation of existing knowledge.
Capable of Handling Difficult Decisions and Problems: An expert system can manage challenging decision-making problems and provide solutions.

Capabilities of an Expert System

Advisor: It can advise a human being on any specific area of expertise.
Provide Decision-Making Capabilities: It supports decision-making in various fields, such as financial decisions or medical science.
Demonstrate a Device: It can demonstrate any new product, including its features, specifications, and usage.
Problem Solving: It possesses problem-solving skills.
Explain a Problem: It can provide a detailed description of an input problem.
Input Interpretation: It can interpret the input provided by the user.
Outcome Prediction: It can predict outcomes.
Diagnosis: An expert system designed for the medical field can diagnose a disease without requiring multiple components, as it contains various integrated medical tools.

Advantages of Expert Systems

Availability: They are readily available due to the mass production of software.
Less Production Costs: The production costs of expert systems are extremely reasonable and affordable.
Speed: They operate at high speed, reducing the amount of work required.
Lower Error Rate: The error rate is much lower than that of human errors.
Low Risks: They can operate in environments that are dangerous for humans.
Stable Response: They avoid movements, tension, and fatigue.

1.4 Limitations

Expert systems also have intrinsic and extrinsic limitations, making a human expert necessary. Some of these limitations include:

Unable to provide a creative response in extraordinary situations.
Errors in the knowledge base can lead to wrong decisions.
The cost of maintaining an expert system is too high.
Every problem is different, so a human expert's solution can also be different and more creative.
Acquiring knowledge for design is much more difficult.
A specific expert system is needed for each domain, which is a significant limitation.
It cannot learn on its own and therefore requires manual updates.

1.5 Applications of Expert Systems

In the Field of Design and Manufacturing

Used widely to design and manufacture physical devices such as camera lenses and automobiles.

In the Field of Knowledge

These systems are mainly used to disseminate relevant knowledge to users. Two popular expert systems in this field are Advisor and Tax Advisor.

In the Financial Field

In financial sectors, expert systems are used to detect possible fraud, identify suspicious activity, and guide bankers on whether to grant loans to businesses.

In Device Diagnosis and Troubleshooting

Expert systems are used in medical diagnosis, which was the first area where these systems were applied.

Planning and Scheduling

Expert systems can also be used to plan and schedule specific tasks to achieve the objectives of those tasks.

2. Composition of Expert Systems

Expert systems essentially consist of three main components:

The Knowledge Base
The Inference Engine
The User Interface

2.1 User Interface

A user interface is the means by which the expert system interacts with a user, taking input queries in a readable format and passing them to the inference engine. Inputs can be provided through dialog boxes, command prompts, forms, or other methods.

The two most important interfaces are:

Consultation Interface: Allows users to obtain advice or solutions from the expert system.
Knowledge Acquisition Interface: Used by human experts or cognitive scientists to update and verify the knowledge base.

After receiving the response from the inference engine, the system displays the output to the user. Some expert systems interact with other computer applications and do not interact directly with a human.

2.2 The Knowledge Base

The knowledge base contains all the facts and rules that a human expert would need to perform their job. It includes the data, knowledge, relationships, principles, and decision rules used by the expert to solve a particular type of problem. The knowledge base stores the accumulated expertise over the years.

It is composed of two subcomponents: the fact base and the rule base.

The Fact Base

The fact base consists of factual knowledge or specific facts of the domain. It includes:

Definitions
Associations
Measurements
Probabilities
Observations
Constraints and assumptions
Statements and assertions

These facts represent what is known about the situation before reasoning begins.

The Rule Base

The rule base specifies the behavior of the system. These rules, written in a structured language, allow the inference engine to calculate or produce a diagnosis. Knowledge related to the problem is "extracted" from domain expertise and expressed in the form of production rules, such as "If condition Then action." These rules dictate how to use facts to reach conclusions.

2.3 The Inference Engine

The inference engine, also known as the control structure, provides a methodology for reasoning about the information in the knowledge base, especially when an answer is not apparent at the time of the initial input. The inference engine manipulates the data encoded in the knowledge base to infer conclusions or predictions.

Inference engines are typically rule-based or logic-based:

Rule-Based Engines: Evaluate rules in a specific order, usually from top to bottom, and apply them to the facts in the database.
Logic-Based Engines: Use symbolic reasoning to perform logical deductions.

Mechanisms of Inference Engines

Forward Chaining (Data-Driven): Starts with the available data and applies rules to infer new information until a goal is reached.
Backward Chaining (Goal-Driven): Begins with a hypothesis or goal and works backward, using rules to determine whether the available data supports the goal.

Conclusion

The integration of expert systems into marketing strategies offers an innovative approach to decision-making. By leveraging AI's capacity to simulate human expertise, companies can enhance their marketing efforts through data-driven insights and predictions. However, the limitations of expert systems, such as their lack of creativity and high maintenance costs, necessitate a balanced approach that includes human judgment alongside technological tools.

Ultimately, expert systems represent a significant advancement in the intersection of technology and business, providing valuable tools for marketers to optimize their strategies in a complex and ever-evolving landscape.