Download - A System to Place Incoming Students in Classes

Transcript
Page 1: A System to Place Incoming Students in Classes

{{

1

A System to Place A System to Place Incoming Students Incoming Students in Classesin Classes

Henry M. Walker, Andrew Hirakawa, Russel SteinbachHenry M. Walker, Andrew Hirakawa, Russel SteinbachGrinnell College, Grinnell, IowaGrinnell College, Grinnell, Iowa

Page 2: A System to Place Incoming Students in Classes

2

Three specific objectives:Three specific objectives:

1.1.Identify most appropriate courseIdentify most appropriate course

2.2.Communicate placement to students, advisorsCommunicate placement to students, advisors

3.3.Publicize interesting courses, recruitPublicize interesting courses, recruit

Problem: Place Problem: Place incoming students in incoming students in CS, math, statisticsCS, math, statistics

Page 3: A System to Place Incoming Students in Classes

3

Placement ApproachesPlacement Approaches1.1. Based on SAT, ACT, similar aptitude testsBased on SAT, ACT, similar aptitude tests

• Often required for admissionOften required for admission

• Tests measure aptitude, not backgroundTests measure aptitude, not background

2.2. Placement testsPlacement tests• Numerous tests commercially availableNumerous tests commercially available

• Logistical issues in administrationLogistical issues in administration

• calibration needed with local courses calibration needed with local courses

3.3. Placement by facultyPlacement by faculty1.1.Can tap faculty expertiseCan tap faculty expertise

• Time consumingTime consuming

• Variation possible from different facultyVariation possible from different faculty

4.4. Use of an expert systemUse of an expert system• Available for incoming, prospective studentsAvailable for incoming, prospective students

• Allows follow-up studies, based on performanceAllows follow-up studies, based on performance

• Needs non-trivial development timeNeeds non-trivial development time3

Page 4: A System to Place Incoming Students in Classes

4

Basic Approach for an Basic Approach for an Expert SystemExpert System

• Use available student dataUse available student data• Standardized tests: SAT, ACT, AP, IBStandardized tests: SAT, ACT, AP, IB

• High school transcriptHigh school transcript

• Base placements on rulesBase placements on rules• Initial rules from faculty experienceInitial rules from faculty experience

• Follow-up studies: courses taken, gradesFollow-up studies: courses taken, grades

• Inference engine applies rules to dataInference engine applies rules to data

• Develop placement letters, Web pagesDevelop placement letters, Web pages

• Develop softwareDevelop software• Version 1 (1993) LISP-based, TMYCINVersion 1 (1993) LISP-based, TMYCIN

• Version 2 (2010) PHP, locally developedVersion 2 (2010) PHP, locally developed4

Page 5: A System to Place Incoming Students in Classes

5

AcknowledgmentsAcknowledgments1.1. 1993 version (LISP based)1993 version (LISP based)

• TMYCIN Inference Engine: Gordon Novak, AI TMYCIN Inference Engine: Gordon Novak, AI Lab, University of Texas at AustinLab, University of Texas at Austin

• Vikram Subramaniam, Ivan SykesVikram Subramaniam, Ivan Sykes

2.2. ConsultantsConsultants• 1993-2004: Eugene Herman, Charles Jepsen1993-2004: Eugene Herman, Charles Jepsen

• 2004-2009: Emily Moore2004-2009: Emily Moore

• 2009-present: Shonda Kuiper, Chris French, 2009-present: Shonda Kuiper, Chris French, Karen ShumanKaren Shuman

• 2010: Barbara Johnson2010: Barbara Johnson

3.3. Current versionCurrent version• Grinnell CS faculty Grinnell CS faculty

• Dean’s Office for summer “MAP” fundingDean’s Office for summer “MAP” funding5

Page 6: A System to Place Incoming Students in Classes

6

Outputs & System Outputs & System OverviewOverview

Our system operates in two primary modes:•Generation of letters to students based on data from the Registrar.

Sample Student LaTeX Letters

2. Web-based, tentative placements to prospective students.

Web Placement Interface

Page 7: A System to Place Incoming Students in Classes

7

Technical OverviewTechnical Overview• Data for incoming studentsData for incoming students

• Files from Registrar’s OfficeFiles from Registrar’s Office

• Storage in MySQL databaseStorage in MySQL database

• Web-based materialsWeb-based materials• Data from HTML forms (not stored in MySQL)Data from HTML forms (not stored in MySQL)

• Programming in PHPProgramming in PHP• Connection with MySQLConnection with MySQL

• Web scriptingWeb scripting

• Associative arrays for rule storageAssociative arrays for rule storage

• Inference engineInference engine• TMYCIN (1993) ==> PHP (2010)TMYCIN (1993) ==> PHP (2010)

• Backtracking algorithm for rule processingBacktracking algorithm for rule processing

• NotesNotes• Formal grammar for rule structureFormal grammar for rule structure

• Rules in LISP/PHP format (for scripting)Rules in LISP/PHP format (for scripting)

• Rules printable in English format (faculty)Rules printable in English format (faculty)

7

Page 8: A System to Place Incoming Students in Classes

8

•Rules composed have three partsRules composed have three parts• NameName

• ConclusionConclusion

• Conditions Conditions

•Examples:Examples:

array("statRule120","TSTAT-PLACE = 208",

array("all","stdscores = unknown",

array("some","statsem >= 2",

"apstat >= 1")));

array("satRealRule10", "satmathreal =field satmath",

array("all", "satmath >= sat2math1",

"sat2math2adj < satmath"));

Rule structureRule structure

Page 9: A System to Place Incoming Students in Classes

9

Rule ConditionsRule Conditions• Each Condition Array has two partsEach Condition Array has two parts

• Quantifier (e.g., some, all)Quantifier (e.g., some, all)

• Conditions to be satisfiedConditions to be satisfied• Simple conditionsSimple conditions

array("satRealRule10", "satmathreal =field satmath", array("all", "satmath >= sat2math1", "sat2math2adj < satmath"));

• Condition arrayCondition array

array("statRule120","TSTAT-PLACE = 208", array("all","stdscores = unknown", array("some","statsem >= 2", "apstat >= 1")));

9

Page 10: A System to Place Incoming Students in Classes

10

Rule ConclusionsRule Conclusions• Two types of conclusionTwo types of conclusion

• Fixed value conclusionsFixed value conclusions "stdscores = unknown"

• Variable valuesVariable values "satmathreal =field satmath"

• Three partsThree parts• FieldField

• OperatorOperator

• ValueValue

10

Page 11: A System to Place Incoming Students in Classes

11

EfficiencyEfficiency

• Short circuit evaluationShort circuit evaluation

• Rules are markedRules are marked

• Minimize SQL queries.Minimize SQL queries.

11

Page 12: A System to Place Incoming Students in Classes

12

Conclusion & Future WorkConclusion & Future WorkFuture Work:•Ever changing curricula both high school and college demand changes in inference rules.•Occasional re-evaluation of student performance versus generated placements.

Conclusion•Automated placements are a viable alternative.•Two experiments support this claim:•One experiment showed individual placements of students by faculty as reliable as nation-wide placement test.•Second experiment showed rule-based system placements were comparable with placements by faculty.