Free MS DOS versions of the MDSX Library of programs with supporting documentation are available download where indicated |
| Notes for each program: | |
| | Program: | As named by originator |
| | Source: | Source of original program (numerical sub-routines etc. updated by MDSX staff) |
| | Data: | Usually: Way and Mode (see Carroll & Arabie) |
| | Model: | As specified by originator's documentation |
| | Transform: | Transformation Function, also called Level of Measurement |
| | Int/External | External (with user-supplied configuration), or internal (derived from input data) analysis |
| | Hierarchy: | Single program, or hierarchy of models/phases |
| | Representation: | How mode/s are represented in the configuration according to the model |
| | TUG: | Section of The User's Guide to Multidimensional Scaling in which the program is discussed |
| | | N.B. All programs are also separately documented in the User Manual |
| | | |
Program name & [source] | Description |
PRELIMINARY Downloads | Information about the MDSX series; Running MDSX programs; Attribution; |
BBDIAM Source: Brusco | Branch and Bound DIAMeter clusteringDATA: 2-Way 1-Mode MODEL: Partition Clustering TRANSFORMATION: Ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Exclusive clusters |
CANDECOMP Bell Labs Downloads | CANonical DECOMPosition DATA: internal analysis of a set of 3- to 7-way data matrix of (dis) similarity matrices, MODEL: Product: Vector, Scalar Products, Factor, Composition TRANSFORMATION: metric/linear. INT/EXTERNAL: I & E HIERARCHY: n REPRESENTATION: Unidimensional Scale values for each way TUG: 7.1.1, 7.2.2 |
CONPAR Brusco | CONcordance PARtitioningDATA: Internal analysis of a set of three-way (2-mode) data matrix consisting of a set of (dis)similarity matricesMODEL: A two part model: 1) Subjects are partitioned into homogeneous clusters, using BBDIAM (q.v.) and an aggregate dissimilarity matrix produced for each cluster.. 2)Each cluster's data rae scaled using MINISSA Distance scalings TRANSFORMATION: Ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: separate scalings |
CORRESPBrier | CORRESPondence analysisDATA: 2-Way 2-mode table (set of profiles) MODEL: Chi-square distance TRANSFORMATION: Ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: N-mode point |
HICLUS Johnson Downloads | HIerarchical CLUStering DATA: two-way (1 mode) (dis) similarity data MODEL: hierarchical clustering scheme (ultrametric : Diameter & Connectedness solutions) TRANSFORMATION: ordinal, non-metric monotonic INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Inclusive partitions / Tree TUG: 6.1.6 |
INDSCAL-S Bell Labs Basic 3-way model Downloads | INdividual Differences SCALing - SymmetricDATA: internal analysis of a three-way (2-mode) data matrix consisting of a set of (dis) similarity matrices MODEL: Weighted Euclidean Distance / product composition TRANSFORMATION: metric/ linear INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Stimulus Points, subject weights TUG: 7.1.1, 7.2, A7.2 |
MDPREF Bell Labs Downloads | MultiDimensional PREFerence Scaling DATA: internal analysis of two-way, 2-mode data of either a set of paired comparisons matrices or a rectangular, row-conditional matrix of ratings/rankings MODEL: Product TRANSFORMATION: linear INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Stimulus Points, subject vectors TUG: 5.3.2, 6.2.2 |
MDSORT Takane | MultiDimensional SORTingDATA: two-way, 2-mode categorical data (set of sortings) MODEL: product TRANSFORMATION: linear INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Point TUG: (Coxon 1999) |
MINI-RSA Roskam Downloads | (Michigan-Israel-Netherlands Integrated = MINI) Rectangular Smallest space Analysis DATA: internal analysis of two-way, 2-mode data in a rectangular (row-conditional) matrix of (preference) rankings or ratings MODEL: Euclidean distance (Unfolding) model TRANSFORMATION: ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Point-point TUG: 5.3.3.1, 6.2.3 |
MINISSA (N) Roskam Basic Nonmetric Model Downloads | [MINI] Smallest Space Analysis (Nijmegen version) DATA: internal analysis of a two-way (1mode) symmetric matrix of (dis) similarities MODEL: Euclidean distance TRANSFORMATION: Ordinal also called non-metric monotonic INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Point TUG: Ch. 3 |
MRSCAL Roskam Downloads | MetRic SCALing DATA: internal analysis of a two-way (1 mode) symmetric matrix of (dis) similarities MODEL: Minkowski r-metric (default Euclidean distance model) TRANSFORMATION: Linear and Log-interval INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Point TUG: 6.1.4 |
PARAMAP Bell Labs | PARAmetric MAPpingDATA: Two-way, 2-mode data (Profiles) MODEL: Smoothness TRANSFORMATION: Continuity INT/EXTERNAL: I, E HIERARCHY: n REPRESENTATION: Points (1 mode) TUG: 5.2.2, 6.1.5 |
PINDIS Roskam Lingoes Downloads | Procrustean INdividual DIfferences Scaling (Hierarchy) DATA: Two-way 2-mode Configuration Co-ordinates MODEL: Hierarchy of Procrustean Fitting models (general distance and general vector) TRANSFORMATIONS: Similarity, then progressively more complex INT/EXTERNAL: I, E HIERARCHY: y (6 models) REPRESENTATION: Point (dist.), Vector TUG: 7.3 - 7.6, A7.1 |
PREFMAP Bell Labs Downloads | PREFerence MAPping (Hierarchy) DATA: external (and quasi-internal) analysis of two-way, 2-mode row- conditional data (usually a preference measure) MODEL: 3 general distance and 1 vector models TRANSFORMATION: Linear, ordinal INT/EXTERNAL: E, I HIERARCHY: y 4 REPRESENTATION: Point, vector TUG: 4.2, 4.4, 6.2.1, 7.5 |
PROCRUSTES Roskam Lingoes Downloads | PROCRUSTEan Similarity (=PINDIS0)DATA: 2-way 2-mode Configuration Co-ordinates MODEL: Euclidean Distance TRANSFORMATION: Similarity (Reflection, Rotation, Uniform re-scaling) INT/EXTERNAL: I,E HIERARCHY: n REPRESENTATION: Point (dist.) TUG: A7.1 |
PRO-FIT Bell Labs Downloads | PROperty FITting DATA: external analysis of a configuration, using 2-way 2-mode rectangular matrix of property ratings or rankings MODEL: scalar products (vector) TRANSFORMATION: either a linear or continuity INT/EXTERNAL: E HIERARCHY: n REPRESENTATION: vector TUG: 5.22, 6.2.1 |
TRISOSCAL Roskam-Prentice Downloads | TRIadic Similarities Ordinal SCALing DATA: internal analysis of a set of triadic (dis) similarity measures MODEL: Minkowski distance TRANSFORMATION: Ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: point TUG: 2.1.4, 6.1.3 |
CONJOINT (Formerly UNICON) Roskam
| CONJOINT measurement (UNICON, ADDIT, MONANOVA)DATA: N-way table MODEL: Simple composition models TRANSFORMATION: Ordinal INT/EXTERNAL: I HIERARCHY: n REPRESENTATION: Unidimensional scale values for each way TUG: 5.3.1, 6.1.8 |
WOMBATS [Coxon-Sykes] Downloads | Work Out Measures Before Attempting To Scale (Utility) DATA: From rectangular raw data matrix Computes one or more measures of dis/similarity , and Outputs in a user-chosen matrix format (for input into MDSX and other programs) MODEL: Dis/similarity Measures TUG: 2.2 |
Programs in the original MDSX (SV3.2) Library not included, but still available: MINICPA (GLR: [MINI] Conditional Proximity Analysis). MVNDS (BL: Shepard’s Maximum Variance Non-Dimensional Scaling) PARAMAP (BL: Carroll’s PARAmetric MAPping) UNICON (GLR: UNIdimensional CONjoint measurement) CONCOR (Arabie: CONvergence of iterated CORrelations: blockmodel analysis through the production of the image matrices) |
