# Data Analyses in Harris, Uruo and Tang (2024) - **Statistical Analysis:** Kevin Tang - **Date:** 15/02/2024 ## Libraries required for this analysis: 'caret','ndl','lattice','rgl','plot3D','corrplot','dplyr','tidyr', 'ggplot2', 'brms', 'reshape2', 'plyr', 'Hmisc', 'dendextend', 'ggraph', 'gtools', 'R.utils' ## Script files contained in this analysis: - `scripts/analyses.html` : Compiled report of all the data analyses and visualisation. It can be regenerated using `analyses.Rmd` and `data` - `scripts/analyses.Rmd` : R script to conduct the data analyses and visualisation ## Data file - `data/duration.tsv`: A tab-delimited file that contains the duration of the window with which edge was measured in seconds. + `word`: the file name of that stimulus + `duration_sec`: the duration of the stimulus in seconds - `data/edge_original.tsv`: A tab-delimited file that contains the edge measurements using Harris and Uruo's frequency bands + `word`: : the file name of that stimulus + The other variables have the following naming format: e.g., `100_5000_raw.int_max` means from the maximum raw intensity value in the 100-5000 Hz frequency band. Only the type `XXX_XXX_raw.int_sd` was used in the paper. + The lower end of the frequency band + The upper end of the frequency band + Type of value: It can be one of the following types + `raw.int`: The raw intensity value + `d1.int`: the first difference value of `raw` + `d2.int`: the second difference value of `raw` + Type of measurement: It can be one of the following types + `min`: minimum value + `max`: maximum value + `avg`: average value + `sd`: the standard deviation value - `data/edge_kingston.tsv`: A tab-delimited file that contains the edge measurements using Liu and Kingston's frequency bands. The variable descriptions are the same as `data/edge_original.tsv` - `data/noise_original.tsv`: A tab-delimited file that contains the noise measurements + `word`: : the file name of that stimulus + `t.min.energy`: the time at which the energy is at the minimum + `noise`: the noise measurement - `data/ndl.varimp.df`: variable importance output - `data/freqband_cortable.csv`: correlation table of frequency bands of `data/edge_original.tsv` and `data/edge_kingston.tsv` - `model_comparison`: It contains the model comparison results. The naming convention of the individual files: + `loo_waic_bf_` + either edge or noise. If edge, `HU.` + the lower end of the frequency band + `.` + the upper end of the frequency band, and if noise, `Noise` + `_` + position 1 (note: B means [, E means ] in the paper) + `_` + position 2 (note: B means [, E means ] in the paper) + `.tsv` + e.g., `loo_waic_bf_HU.100.2000_VBCV_VCCV.tsv` contains the model comparison of the edge values at the frequency band 100 to 2000 Hz for the two position V[CV and VC:V. ## Image files - `imgs`: Directory contains all the plots reported in the paper (and others that were not reported) in both .pdf and .png formats