Compare lists of texts, factors, or numerical values to measure their similarity. Use cases include comparing responses from multiple raters, evaluating model outputs, and assessing data consistency.
Installation
From CRAN:
# install.packages("pak")
pak::pak("samesies")Development version:
pak::pak("dylanpieper/samesies")Basic Usage
samesies provides three main functions for measuring similarity:
same_text()
Compare similarity between multiple lists of character strings:
library(samesies)
r1 <- list("R is a statistical computing software",
"R enables grammar of graphics using ggplot2",
"R supports advanced statistical models")
r2 <- list("R is a full-stack programming language",
"R enables advanced data visualizations",
"R supports machine learning algorithms")
tex <- same_text(r1, r2)
#> ✔ Computed osa scores for "r1_r2" [mean: 0.43]
#> ✔ Computed lv scores for "r1_r2" [mean: 0.43]
#> ✔ Computed dl scores for "r1_r2" [mean: 0.43]
#> ✔ Computed hamming scores for "r1_r2" [mean: 0.123]
#> ✔ Computed lcs scores for "r1_r2" [mean: 0.061]
#> ✔ Computed qgram scores for "r1_r2" [mean: 0.682]
#> ✔ Computed cosine scores for "r1_r2" [mean: 0.771]
#> ✔ Computed jaccard scores for "r1_r2" [mean: 0.735]
#> ✔ Computed jw scores for "r1_r2" [mean: 0.818]
#> ✔ Computed soundex scores for "r1_r2" [mean: 0.667]Methods available via stringdist (e.g., method = "osa"):
-
Edit Distance Methods
- osa, lv, dl
-
Token-Based Similarity
- hamming, lcs, qgram, cosine, jaccard
-
Phonetic Methods
- jw, soundex
same_factor()
Compare similarity between multiple lists of categorical data:
f1 <- list("R", "R", "Python")
f2 <- list("R", "Python", "R")
fct <- same_factor(f1, f2)
#> ℹ Skipping 'order' method as factor levels are not ordered
#> ✔ Computed exact scores for "f1_f2" [mean: 0.333]Compare similarity based on ordered factors:
of1 <- list("High School", "Bachelor's", "Master's", "PhD")
of2 <- list("Bachelor's", "High School", "PhD", "Master's")
edu_comparison <- same_factor(of1, of2,
levels = c("High School", "Bachelor's", "Master's", "PhD"))
fct_ordered <- average_similarity(edu_comparison)
#> ✔ Computed exact scores for "of1_of2" [mean: 0]
#> ✔ Computed order scores for "of1_of2" [mean: 0.667]Methods available (e.g., method = "exact"):
- exact: Exact matching
- order: Distances across ordered factor levels
same_number()
Compare similarity between multiple lists of numeric values:
n1 <- list(1, 2, 3)
n2 <- list(1, 2.1, 3.2)
num <- same_number(n1, n2)
#> ✔ Computed exact scores for "n1_n2" [mean: 0.333]
#> ✔ Computed raw scores for "n1_n2" [mean: 0.1]
#> ✔ Computed exp scores for "n1_n2" [mean: 0.908]
#> ✔ Computed percent scores for "n1_n2" [mean: 0.963]
#> ✔ Computed normalized scores for "n1_n2" [mean: 0.955]
#> ✔ Computed fuzzy scores for "n1_n2" [mean: 0.977]Methods available (e.g., method = "exact"):
- exact: Exact matching
- raw: Absolute difference
- exp: Exponential decay on the absolute difference
- pct_diff: Percentage difference
- normalized: Normalized difference
-
fuzzy: Intelligent threshold-based matching that adapts to data scale:
- Uses two tolerance thresholds: absolute (calculated) and relative (default 2%)
- Calculation of the absolute threshold averages data variability (10% of standard deviation), magnitude (0.5% of mean absolute values), and range (1% of value range)
- Effective threshold is the larger of these two values
- Perfect matches (score = 1.0) when difference ≤ effective threshold
- Scores decrease smoothly as differences exceed the threshold
List Support
Named Lists
For more control over results, you can use named lists to specify custom labels for comparisons:
result_number <- same_number("baseline" = n1, "treatment" = n2, method = "fuzzy")
#> ✔ Computed fuzzy scores for "baseline_treatment" [mean: 0.978]More Lists
When you input more than two lists, samesies computes pairwise comparisons across all lists:
r1 <- list("Statistical computing", "Data visualization", "Machine learning")
r2 <- list("Statistical software", "Data plotting", "ML algorithms")
r3 <- list("Statistical analysis", "Data graphics", "AI models")
multi_text <- same_text(r1, r2, r3, method = "cosine")
#> ✔ Computed cosine scores for "r1_r2" [mean: 0.717]
#> ✔ Computed cosine scores for "r1_r3" [mean: 0.607]
#> ✔ Computed cosine scores for "r2_r3" [mean: 0.68]Nested Lists
Compare nested lists with identical structure and names:
nest1 <- list(
group_a = list("Good", "Very Good", "Fair"),
group_b = list("Excellent", "Good", "Poor")
)
nest2 <- list(
group_a = list("Very Good", "Good", "Fair"),
group_b = list("Good", "Excellent", "Fair")
)
nested_result <- same_text(nest1, nest2, method = "jw")
#> ✔ Computed jw scores for "nest1_nest2" [mean: 0.25]Methods
All three functions return similar objects that support the following methods:
result <- same_text(r1, r2, method = "cosine")
# Print detailed results
print(result)
# Summarize results
summary(result)
#> method pair avg_score
#> cosine r1_r2 0.771
# Get average similarity scores
average_similarity(result)
#> cosine
#> 0.771
# Get pair-wise averages
pair_averages(result)
#> method pair avg_score
#> 1 cosine r1_r2 0.771Accessing Object Data
The package uses S3 objects, allowing access to the underlying data using $:
result <- same_text(r1, r2, method = "cosine")
# Access similarity scores
result$scores
#> $cosine
#> $cosine$r1_r2
#> R is a statistical computing software
#> 0.7917448
#> R enables grammar of graphics using ggplot2
#> 0.7027498
#> R supports advanced statistical models
#> 0.8197365
# Access methods used
result$methods
#> [1] "cosine"
# Access list names
result$list_names
#> [1] "r1" "r2"Available components:
-
scores: A list of similarity scores for each method and comparison pair -
summary: A list of statistical summaries for each method and comparison pair -
methods: The similarity methods used in the analysis -
list_names: Names of the input lists -
raw_values: The original input values -
digits: Number of decimal places for rounding results in output
Credits
The Spiderman image in the hex logo is fan art created by the Reddit user WistlerR15.