codecountR: Counting Codes in a Text and Preparing Data for Analysis
Data analysis often requires coding, especially when data are collected through interviews, observations, or questionnaires. As a result, code counting and data preparation are essential steps in the analysis process. Analysts may need to count the codes in a text (Tokenization, counting of pre-established codes, computing the co-occurrence matrix by line) and prepare the data (e.g., min-max normalization, Z-score, robust scaling, Box-Cox transformation, and non-parametric bootstrap). For the Box-Cox transformation (Box & Cox, 1964, <https://www.jstor.org/stable/2984418>), the optimal Lambda is determined using the log-likelihood method. Non-parametric bootstrap involves randomly sampling data with replacement. Two random number generators are also integrated: a Lehmer congruential generator for uniform distribution and a Box-Muller generator for normal distribution. Package for educational purposes.
| Version: |
0.0.4.8 |
| Imports: |
stats |
| Suggests: |
knitr, rmarkdown |
| Published: |
2025-03-01 |
| Author: |
Philippe Cohard [aut, cre] |
| Maintainer: |
Philippe Cohard <p.cohard at laposte.net> |
| License: |
GPL-3 |
| NeedsCompilation: |
no |
| CRAN checks: |
codecountR results |
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=codecountR
to link to this page.