WordStat 5.1 Computer Assisted Text Analysis

LIST OF FEATURES

TEXT PROCESSING CAPABILITIES

• Content analysis on short alphanumeric variable (up to 255 characters) and longer ANSI or RTF document (several mb).
• Dictionary moderated lemmatization and stemming (English, French, Italian and Spanish; contact us for other languages).
• Ability to call external text pre-processing EXE or DLL (sample English porter stemmer and n-grams transformation are include)
• Optional exclusion of pronouns, conjunctions, etc, by the use of user-defined exclusion lists (or stop list).
• Categorization of words or phrases using existing or user-defined dictionaries.
• Word categorization based on Boolean (AND, OR, NOT) and proximity rules (NEAR, AFTER, BEFORE)
• Word and phrase substitution and scoring using wildcards and weighting.
• Frequency analysis on keywords, phrases, derived categories or concepts, or user-defined codes entered manually within a text.
• Interactive development and easy maintenance of hierarchical dictionaries, taxonomies, or categorization schema.
• Drag and drop editor for easy assignments of words, phrases into categories!
• Ability to restrict the analysis to specific portions of a text or to exclude comments and annotations.
• Ability to perform an analysis on a random sample of cases.
• Integrated spell-checking with support for different languages such as English, French, Spanish, etc.
• Integrated thesaurus (English only) to assist the creation of taxonomies and comprehensive categorization schemas.
• Powerful case filtering on any numeric or alphanumeric field and on code occurrence (with AND, OR, and NOT boolean operators)
• Prints presentation quality tables
• Imports MS Word, WordPerfect, RTF and HTML.
• Exports any table to Excel, ASCII, Tab separated or comma separated value files, or HTML files.
• Flexible keyword highlighting (the text editor can display all categories using different colors).

UNIVARIATE KEYWORD FREQUENCY ANALYSIS

• Univariate word frequency analysis (word or category count and record occurrence).
• Word x word co-occurrence matrix.
• Word x case data matrix.
• Integrated multidimensional scaling with 2D and 3D maps.
• Proximity plot.

FEATURE EXTRACTION

• Vocabulary finder extracts technical terms, product and company names as well as common misspellings.
• Phrase finder allows one to easily identify recurring phrases and expressions

NORM CREATION AND COMPARISON

• Ability to create norm files based on frequency analysis of words or content categories.
• Comparison of obtained frequencies to previously saved norm files.

KEYWORD RETRIEVAL FUNCTION

• A powerful keyword retrieval function allows identification of text units (documents, paragraph or sentences) containing one keyword or a combination of keywords with optional filtering of cases.
• Ability to attach QDA Miner codes to retrieved segments.
• Retrieved segments may be exported to disk in tabular format (Excel or delimited text files) or as text reports (Rich Text Format).

KEYWORD CO-OCCURRENCE ANALYSIS

• Integrated clustering and dendrogram display of keyword co-occurrence.
• First- and second-order proximity analysis.
• Proximity plot to easily identify all keywords that co-occurs with a target keyword.
• 2D and 3D multidimensional scaling on either joint frequency or co-occurrence of words or categories.
• Flexible keyword co-occurrence criteria (within a case, a sentence, a paragraph, a window of n words, a user-defined segment) as well as clustering methods (first- and second-order proximity, choice of similarity measures).
• Easy text retrieval from dendrogram or proximity plots.

ANALYSIS OF CASE OR DOCUMENT SIMILARITY

• Hierarchical clustering, multidimensional scaling and proximity plot may be used to explore the similarity between documents or cases.

MULTIPLE RESPONSES AND COMPARISONS

• Can perform univariate frequency analysis and crosstabulation on information stored in several alphanumeric fields (memo or string variables).
• Comparison of keyword occurrence between different fields.
• Computes inter-raters agreement measures (pct. of agreement, Cohen's Kappa, Scott's Pi, Krippendorff's R and r-bar, free marginal) based on codes manually entered in different variables.

BIVARIATE COMPARISONS BETWEEN SUBGROUPS

• Bivariate comparison between any textual field and any nominal or ordinal variable (such as the sex of the respondent, specific subgroups, years of publication, etc.).
• Choice between 11 different association measures to assess the relationship between word occurrence and nominal or ordinal variables (Chi-square, Likelihood ratio, Tau-a, Tau-b, Tau-c, symmetric Somers' D, asymmetric Somers' Dxy and Dyx, Gamma, Person's R, Spearman's Rho)
• Computation statistics on either absolute or relative frequency
• Ability to sort matrix in alphabetic order of words, by word frequency or word occurrence, on the obtained statistics or on its probability.
• Visually compare items between subgroups using bar charts and line charts.

• Correspondence analysis (statistics, 2D & 3D joint plots). This new feature is accessible from the crosstab page and allows one to see graphically the relationship between nominal variables and codes resulting from a content analysis.
• Heatmap plot (with dual-clustering of keywords and variables)

AUTOMATED TEXT CLASSIFICATION

• Machine learning algorithms (Naive Bayes and K-Nearest Neighbors) for document classification.
• Flexible feature selection for automatic selection of best subsets of attributes.
• Numerous validation methods (leave-but-one, n-fold crossvalidation, split sample).
• Experimentation module allows easy comparison of predictive models and fine-tuning of classification models.
• Classification models may be saved to disk and applied later using either a standalone document classification utility program, a command line program or a programming library . Note: The command line and the programming library are part of WordStat Software Developer's kit (SDK) which is sold separately.

KEYWORD-IN-CONTEXT (KWIC)

• Ability to display a KWIC table to examine the textual context of a word, word pattern, or category.
• Ability to sort the table on any independent (numeric) variables.
• Ability to jump from a KWIC keyword to the textual variable in order to view or edit the original text.
• KWIC list can be saved in data files for further processing.
• Customizable KWIC display (paragraph, sentence or user defined segment).
• Concordance report (displays all hits as a list of paragraphs, sentences or user defined segments)

FULL INTEGRATION WITH A STATISTICAL SOFTWARE

• Alphanumeric variables can be stored in the same file as all other numeric variables.
• Variable selection, statistical analysis and content analysis are performed within the same application program.
• Matrix outputs are automatically added to existing statistical outputs.
• New variables representing occurrence of words, keywords or concepts can be added to the existing data file or exported to a new data file in order to be submitted to further statistical analysis (such as cluster analysis on words or cases, principal coordinate analysis, correspondence analysis, multiple regression, etc.).
• Data can be imported from and exported to different file format including dBase, Paradox, Excel, Quattro Pro, Lotus 1-2-3, SPSS for DOS, SPSS for Windows, comma or tab separated text files, etc.
• Ability to perform numeric and alphanumeric transformation or to apply filters on records of the data file to restrict the analysis to specific subgroups. .

UTILITY PROGRAMS

• Dictionary building assistant to find related words (synonyms, antonyms, holonyms, meronyms, hypernyms, hyponyms) in a WordNet based thesaurus (English only). (100,000 synonyms, 120,000 root words)

• WS Document Classifier, a small standalone application to apply previously saved categorization and classification models to external documents.

• WSTOOLS - Utility program to easily import documents of any size into Simstat database files. Various file formats may be directly imported such as:
  • Plain text (with optional DOS ASCII to Windows ANSI conversion)
  • HTML (with or without removal of HTML tags)
  • RTF
  • MS Word *
  • WordPerfect *
  • Excel *             (* The support for specific versions of these file formats may differ depending on your windows version and configuration.)

• Optional removal of leading and trailing spaced and hard returns.
• Extraction of numeric and alphanumeric variables from documents.
• Extraction options may be saved on disk and later retrieved.
• Documents may be stored as plain ANSI text or as RTF documents.