Expert Systems: Potential Management Aids

Farmers and ranchers are often envious of large businesses that can afford consultants to sift through information and help them make informed management decisions. A new information technology called expert systems (ES), or knowledge-based systems, offers the potential of bringing the consultant to the farm through microcomputers. ES are computer programs (software) that provide expertise to address a specific question and draw conclusions equal in quality to those one would expect from a human expert. ES also provide reasoning behind conclusions based on rules incorporated into the software. By contrast, decision aid and accounting software now produce only numbers, leaving interpretation to the user. ES provide expertise, through use of rules in the software, to interpret the results.

ES can help producers make decisions in an increasingly complex environment that requires a great deal of specialized information. Much of the present information available in agriculture does not tell decisionmakers what they need to know. Sometimes referred to as "information overload," what is provided is simply a great deal of unusable data and information. ES can help farm managers sort through the multitude of on- and off-farm data and information, determining what is useful. Although no new knowledge is generated by ES, they do provide access to existing knowledge for the decisionmaker who needs expert help.

The use of ES applications in agriculture is best illustrated by three examples of ES presently available to decisionmakers. These three ES are diagnostic systems dealing with crop variety selection, disease diagnosis, and financial analysis.

Wheat Variety Selection

WHEAT WIZ is an expert system developed at Kansas State University for selection of hard red wheat varieties. The data base contains pedigree and release information, disease and insect resistance, maturity and winter hardiness rating, and relative yield on 180 hard red winter wheat varieties. The software can provide information about a certain variety, identify varieties that have the best resistance to specific pests, or provide a list of adapted varieties based on user-specified field location, pest problems, and cultural practices. WHEAT WIZ recommends varieties fora particular field using the decision process that would be expected from a wheat specialist. This microcomputer program is available through the Kansas Cooperative Extension Service.

Soybean Disease Diagnosis

Soybean Disease Diagnosis is an expert system developed at the University of Illinois by J. B. Sinclair and Ryszard S. Michalske to diagnose soybean diseases common in Illinois. The system identifies the disease based on a user's answers to specific questions about the diseased soybean plant and its growing environment. The program expresses its final opinion in terms of the degree of confidence that the plant has a specific disease. The degree of confidence is based on the rules that plant pathologists have in the program. The program ends by providing prescriptive information on chemical and natural control for diseases. This expert system is a useful tool for Illinois soybean producers and is distributed to them through the Illinois Cooperative Extension Service.

Agricultural Financial Analysis Expert Systems (AFAES)

AFAES is a set of software designed to facilitate organization and analysis of farm and ranch financial data. I developed these programs with Kedric Karkosh and Clark Osborne. AFAES include software to be used in developing and analyzing data for agricultural financial statements, along with expert systems as it performs diagnostic analysis of the farm or ranch business' financial condition and performance. The analysis is similar to what one would expect from an agricultural financial analysis expert. This software can be used by producers, agricultural lenders, accountants, and educators.

The ES presently evaluate a farm or ranch (1) operating year performance, (2) financial condition, and (3) ability to support operating capital debt. Analysis is based on historical, current, and projected financial data. Additional ES are under development to evaluate the feasibility of long-term capital investment based on financial projections.

Participants in Development. AFAES software was developed by the Texas Agricultural Experiment Station, Texas A & M University. Cooperators in the development and evaluation of AFAES software include the Farm Credit Bank of Texas, producers in Texas, the Texas Agricultural Extension Service, private lenders, and two national task forces. USDA's Federal Extension Service supported the National Task Force with 35 members from 12 States including land-grant professionals, lenders, and software vendors. Development efforts are also closely coordinated with the American Bankers Association's National Farm Financial Standards task force that is developing accounting and financial reporting standards for agriculture. Developing ES for national efforts will help standardize financial reporting and analysis, making agricultural borrowers more competitive with nonagricultural businesses that have standardized reporting procedures.

Diagnostic Financial Analysis and Explanation. AFAES software facilitates the organization of information for current, historical, and projected balance sheets and accrual income statements. These data are then processed to generate specific financial performance ratios and trends. The expert system's rules then use this information to derive diagnostic evaluation of the business and to explain reasons for conclusions.

An example of the kind of information included in the results of an AFAES analysis is shown in the table. This example uses a dairy farm. The financial ratios and measures would be part of the report, along with a graphic presentation and another table (not shown) that defines the ratios and the acceptable range of values for each ratio. For example, the graphic results (not shown) indicated that the dairy farm had a slightly favorable financial performance.

The program explains each one of the factors which were evaluated. Examples of the interpretation of financial factors provided by the program are given below.

During the operating year analyzed, the firm had a positive cash-flow of $41,014.

A return on farm assets of 5 percent puts the firm in an acceptable profitability position. Each dollar of farm assets is generating $.05 of net income. Profitability should be monitored to insure that it increases.