ALA-Reader (analysis of lexical aggregates) version 1.01

(v1.0, May 2004; v1.01, March 2005) by Roy B. Clariana (RClariana@psu.edu)

Here is a software tool that can translate written text summaries directly into proximity files (prx) that can be analyzed by Pathfinder KNOT. It also generates text proposition files that can be imported by CMAP Tools to automatically form concept maps from the text. It should be of use to researchers who want to visualize "text" for various instructional and research-related reasons. Also it should work with different languages.

ALA-Reader contains a rudimentary scoring system. Essentially, this tool converts the written summary into a cognitive map and then scores the cognitive map using an approach that we developed for scoring concept maps. The "score" produced is percent agreement with an expert referent. As I narrow down what algorithms work, then I plan to release updated versions periodically.

Download ALAReader.exe (a 1.3 Mb download) and the practice files into the same folder. Note, the password is: peace

Launch ALA-Reader by double-clicking on it. It only has 4 buttons, "MARK (S)", "MARK (L)", "TERMS", and "QUIT", so it is fairly easy to use. After you click "MARK (S) or MARK (L)", it will ask you to type in the file name. Be precise when typing or you will get a WARNING: FILE NOT FOUND. If you miss-type, just click MARK again and type the file name correctly.

ALA-Reader creates a folder named "PRX Data Folder" that contains the prx files (you need Schanveldt's KNOT software to analyze these prx files). It also produces a file named report.txt that has the rudimentary score agreement data. We have completed one pilot that suggests that the scores may be adequate for some purposes (read CMC_2004.doc and CMC_2006.htm)

Download the following practice text (essay) files into the same folder as ALA-Reader. Try it with these before trying your own texts.

  1. terms.txt You open this file and add up to 30 most important terms, see the information below on the required file structure for terms.txt (this version of ALA handles up to two synonyms and metonyms for each term)
  2. expert1.txt (add an expert's text that is used to judge the student essays)
  3. expert2.txt (add another referent text that is used to judge the student essays, for example the best student text)
  4. essay01.txt (the first student's text)
  5. essay02.txt (the second student's text, etc.)
  6. essay03.txt
  7. essay04.txt
  8. essay05.txt

Unfortunately, the text (essay) files must follow this format (I'll add sentence parsing in later versions):
Line 1: Title
Line 2: Students name or ID
Line 3: (Blank line)
Line 4: Sentence 1 (notice that each sentence is one line long)
Line 5: Sentence 2
etc. (not to exceed 30 sentences)
Note: Open any text (essay) to see an example of this required format.

Ok, now for a few limitations.

Important - You must set up the terms.txt file correctly.

  1. You can have up to 30 terms, one term per line

  2. Terms that require more than a word (i.e., mitral valve) must be typed in with an underscore rather than a space (e.g., mitral_valve)

  3. Each line has four forms of that term separated by spaces. The first form is standard language, and the other three are shortened forms used for pattern matching. The four include:

Here is an example of terms.txt
lungs lung
membranes membran
mitral_valve mitral bicusp atrioventricular
myocardium myocard

 

Contact me if you need help or have experimental results to report.
I will keep a list of links to Research and Application of this Utility.

...Roy (email: RClariana@psu.edu)

T software by averaging together the