Digital Commons Network^™

Application Of Information Theory To The Construction Of Efficient Decision Trees, Carlos R.P. Hartmann, Pramod Varshney, Kishan Mehrotra, Carl L. Gerberich

Electrical Engineering and Computer Science - Technical Reports

This paper treats the problem of conversion of decision tables to decision trees. In most cases, the construction of optimal decision trees is an NP-complete problem and, therefore, a heuristic approach to this problem is necessary. In our heuristic approach, we apply information theoretic concepts to construct efficient decision trees for decision tables which may include “don’t-care” entries. In contrast to most of the existing heuristic algorithms, our algorithm is systematic and has a sound theoretical justification. The algorithm has low design complexity and yet provides us with near-optimal decision trees.

Algorithms For Pipe Network Analysis And Their Reliability, Don J. Wood

KWRRI Research Reports

Algorithms for analyzing steady state flow conditions in pipe networks are developed for general applications. The algorithms are based on both loop equations expressed in terms of unknown flowrates and node equations expressed in terms of unknown grades. Five methods, which represent those in significant use today, are presented. An example pipe network is analyzed to illustrate the application of the various algorithms. The various assumptions required for the different methods are presented and the methods are compared within a common framework.

The reliabilities of these commonly employed algorithms for pipe network analysis are investigated by analyzing a large number …

Recommendations For Master's Level Programs In Computer Science: A Report Of The Acm Curriculum Committee On Computer Science, Kenneth I. Magel, Richard H. Austing, Alfs Berztiss, Gerald L. Engel, John Wesley Hamblen, A. A.J. Hoffmann, Robert Mathis

Computer Science Faculty Research & Creative Works

The ACM Committee on Curriculum in Computer Science has spent two years investigating master's degree programs in Computer Science. This report contains the conclusions of that effort. Recommendations are made concerning the form, entrance requirements, possible courses, staffing levels, intent, library resources, and computing resources required for an academic, professional, or specialized master's degree. These recommendations specify minimum requirements which should be met by any master's programs. The Committee believes that the details of a particular master's program should be determined and continually updated by the faculty involved. A single or a small number of model programs are not as …

Regular Expressions In A Program Complexity Metric, Kenneth I. Magel

Computer Science Faculty Research & Creative Works

No abstract provided.

A Complexity Measure Based On Nesting Level, Warren A. Harrison, Kenneth I. Magel

Computer Science Faculty Research & Creative Works

No abstract provided.

A Suggested Course In Introductory Computer Programming, Warren A. Harrison, Kenneth I. Magel

Computer Science Faculty Research & Creative Works

Introductory programming courses have long been a popular topic of discussion. Often it is either the only computer course a student takes, or it is the foundation upon which all further training in computer science is built. The usual goal of such a course is to introduce the student to the use of a computer to solve simple problems in his or her particular discipline. Generally, the method of presenting the material may be separated into two distinct schools of thought, which we have termed the "Black Box School of Thought" and the "White Box School of Thought". In this …

The Double Eigenvalue Problem; Including Numerical Solutions, John Gregory, Ralph W. Wilkerson

Computer Science Faculty Research & Creative Works

No abstract provided.

The Coronary Distribution Zone : Mathematical Modelling Of The Left Ventricular Epicardial Surface For Area Calculation, Three-Dimensional Rotaton And Display, S. Matts E. Sjölander

Loma Linda University Electronic Theses, Dissertations & Projects

A computer model of the left ventricular (LV) epicardial surface was developed from Jang’s truncated ellipsoidal model, which has a minor to major axis ratio of 6:10. The model was applied in calculating the area of a segment of the myocardium, jeopardized by an impending occlusion of its associated artery. The endangered segment of the LV muscle was also visualized in different projections by rotations in three space and graphical displays simulating the three dimensions.

The area formula for the surface of the ideal model , a truncated prolate spheroid, was derived and used in producing maps for manual calculations …

The Role Of Technological Change, Robert U. Ayres, Steven M. Miller

Research Collection School Of Computing and Information Systems

We develop an optimal growth model that includes several important new features. First, technological change is endogenously related to the growth of 'knowledge.' Investment may be directed either towards physical capital or knowledge (or both). Knowledge becomes an effective substitute for scarce resources by increasing the technical efficiency of resource utilization both for consumption and in capital. Nevertheless, a finite quantity resource must be embodied in capital and a finite flow is required for depreciation. Thus, there is an upper limit to technical efficiency and economic growth is thus ultimately limited by the availability of renewable resources. For a simple …

Bounds On The Number Of Samples Needed For Neural Learning, Kishan G. Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

Electrical Engineering and Computer Science - Technical Reports

This paper addresses the relationship between the number of hidden layer nodes in a neural network, the complexity of a multi-class discrimination problem, and the number of samples needed for effective learning. Bounds are given for the latter. We show that Ω(min(d,n).M) boundary samples are required for successful classification of M clusters of samples using a 2 hidden layer neural network with d-dimensional inputs and n nodes in the first hidden layer.

Computing Cyclic List Structures, F. Lockwood Morris, Jennifer Schwarz

Electrical Engineering and Computer Science - Technical Reports

It is argued that list structures containing cycles are useful and unobjectionable Lisp entities. If this is so, it is desirable to have a means of computing them less foreign to the equational-definition style characteristic of Lisp than are the list-structure-altering primitives rplaca and rplacd. A notion is developed of a reasonable system of mutually recursive equations, guaranteed to have a unique solution in list structures. The notion is given in terms of the computations invoked by the equations, without reference to the forms of expressions appearing in them. A variety of programming examples are presented, including a curious implementation …

An Information Theoretic Approach To The Construction Of Efficient Decision Trees, Jaime M. De Faria, Carlos R.P. Hartmann, Carl L. Gerberich, Pramod Varshney

Electrical Engineering and Computer Science - Technical Reports

This paper treats the problem of construction of efficient decision trees. Construction of optimal decision trees is an NP-complete problem and, therefore, a heuristic approach for the design of efficient decision trees is considered. The approach is based on information theoretic concepts and the proposed algorithm provides us with a simple procedure for the construction of near-optimal decision trees.

A Software Science Analysis Of Cobol Programs, V. Y. Shen, Herbert E. Dunsmore

Department of Computer Science Technical Reports

No abstract provided.

A Cooperative Education Program For Technology Majors, Joel Lee Sawyer

College of Graduate Studies: Theses & Dissertations (1964–2006)

Cooperative Education is a method of higher education that combines academic work with practical on-the-job experience. The goal is to prepare students for career opportunities through a balance of theory and related employment with a business or industry. "Thus, the student is provided a unique experience. . . Work that will be profitable to both the student and employer; and experience that will enhance the student's knowledge, personal development and professional preparation" (3, p. 2).

A need has been established to initiate a Cooperative Education Program for the Division of Technology at Georgia Southern College. This study was devoted to …

Correction To "Redundancy Optimization Of General Systems", H. Sivaramakrishnan, Arcot Desai Narasimhalu

Research Collection School Of Computing and Information Systems

Reader Aids-

Purpose: Report a correction

Special math needed: Probability

Results useful to: Reliability Theoreticians

Efficient Employment Of Cohorts Of Labor In The U.S. Economy: An Illustration Of A Method, A. E. Boardman, Steven M. Miller, A. P. Schinnar

Research Collection School Of Computing and Information Systems

This paper presents a model for the analysis of efficient labor force participation in the U.S. economy. Ridge regression estimates of the elasticities of cohorts of labor, classified by sex and occupation, are used in conjunction with wage data to illustrate the derivation of efficient allocation of labor cohorts in five economic sectors. These efficient constructs are compared with actual census data for 1960 and 1970. The results, while tentative, show a trend toward more efficient utilization of labor and greater participation by women in the work force.

Hash-Binary Search: A Fast Technique For Searching An English Spelling Dictionary, Douglas E. Comer, Vincent Y. Shen

Department of Computer Science Technical Reports

No abstract provided.

A Proof Procedure For Quantifier-Free Non-Clausal First Order Logic, Neil V. Murray

Electrical Engineering and Computer Science - Technical Reports

A proof procedure is described which operates on formulas of the predicate calculus which are quantifier-free. The procedure, which involves a single inference rule called NC-resolution, is shown to be complete. Completeness is also obtained for a simple restriction on the rule’s application. Examples are given using NC-resolution not only for synthesis of a logic program from its specification, but for execution of a program specification in its original form.

A Mechanism For Specifying Parallel Procedures, James Oliver Smith Jr.

Masters Theses

"This thesis presents a mechanism that will provide a semantic and syntactic environment for expressing parallel procedures. This mechanism will be referred to as the Parallel Machine. The parallelism of the Parallel Machine is accomplished through an arbitrary number of specialized computing elements that each implement a single command called a Simultaneous Command. Each Simultaneous Command is capable of determining the occasions for its action by observing conditions within the Parallel Machine. There is no restriction on the number of Simultaneous Commands that can be defined or be simultaneously active. Within a parallel procedure expressed as a set of Simultaneous …

A Graded Bibliography On Macro Systems And Extensible Languages, John R. Metzner

Computer Science Faculty Research & Creative Works

No abstract provided.

A Simple Experiment In Top-Down Design, Douglas E. Comer, M. H. Halstead

Department of Computer Science Technical Reports

No abstract provided.

An Inductive Inference Bibliography, Carl H. Smith

Department of Computer Science Technical Reports

No abstract provided.

A Rapid Algorithm For Reliability Optimization Of Parallel Redundant Systems, Arcot Desai Narasimhalu, H. Sivaramakrishnan

Research Collection School Of Computing and Information Systems

A rapid method is proposed for optimization of reliability of multiconstraint parallel redundant systems. The constraints need not be linear. This method provides good starting values, which are close to the boundary of the feasible region, for the number of redundant units in each subsystem. No proof has been presented to establish the optimality obtained by this method. Yet for examples tried out this method provides optimal or near optimal solutions.

On List Structures And Their Use In The Programming Of Unification, F. Lockwood Morris

Electrical Engineering and Computer Science - Technical Reports

The notion of list structure is discussed, and a new construct is introduced into LISP which permits the computation of list structures containing cycles without recourse to operations which alter existing structures. It is shown that list structures can be used to represent both finite and “rational” infinite terms. Substitutions (generalized for a term algebra which includes infinite terms) are discussed, “tables” are introduced as an abstract data type, and two methods of representing substitutions by tables, together with their interrelation, are considered. Concise programs are given for a succession of forms of Robinson’s unification algorithm, including one which operates …

Quantitative Estimates Of Debugging Requirements, Linda M. Ottenstein

Department of Computer Science Technical Reports

No abstract provided.

Terminal Talk - The Wofford Connection - June 1978, Wofford College Computer Center

Terminal Talk

No abstract provided.

Geometric Problems With Application To Hashing, Douglas E. Comer, Michael J. O'Donnell

Department of Computer Science Technical Reports

No abstract provided.

Map: A Pascal Macro Preprocessor For Large Program Development, Douglas E. Comer

Department of Computer Science Technical Reports

No abstract provided.

Terminal Talk - The Wofford Connection - February 1978, Wofford College Computer Center

Terminal Talk

No abstract provided.

A Technique To Calculate Complex Electromagnetic Fields By Using The Finite Element Method, Davood Asgharian

Dissertations and Theses

A computer program based on Maxwell's equations is developed to calculate two-dimensional complex potentials by the Finite Element Method. This study offers a solution to a complex continuum problem by allowing a subdivision into a series of simple interrelated problems. The region of interest is divided into triangular elements. For each node in the grid, the Finite Element Method is used to set up an equation for the potential as a function of those of the surrounding nodes. All these equations are solved by the Gaussian Elimination Method. For increased accuracy this method requires a high degree of division of …