dataSDA

Datasets and Basic Statistics for Symbolic Data Analysis

Overview

dataSDA collects a diverse range of symbolic data and offers a comprehensive set of functions that facilitate the conversion of traditional data into the symbolic data format. It supports reading, writing, and conversion of symbolic data in diverse formats, as well as computing descriptive statistics of symbolic variables.

Installation

From GitHub

# install.packages("devtools")
devtools::install_github("hanmingwu1103/dataSDA")

From source

Download the latest release from the Releases page, then:

# Source package (all platforms)
install.packages("dataSDA_0.2.5.tar.gz", repos = NULL, type = "source")

# Binary package (Windows)
install.packages("dataSDA_0.2.5.zip", repos = NULL, type = "win.binary")

Features

Descriptive Statistics

Interval-valued data (`int_*`)

Compute mean, variance, covariance, and correlation for interval-valued data with 8 methods: CM, VM, QM, SE, FV, EJD, GQ, SPT.

library(dataSDA)
data(mushroom.int)

int_mean(mushroom.int, var_name = "Pileus.Cap.Width")
int_var(mushroom.int, var_name = c("Stipe.Length", "Stipe.Thickness"), method = c("CM", "FV", "EJD"))

int_cov(mushroom.int, var_name1 = "Pileus.Cap.Width",
        var_name2 = c("Stipe.Length", "Stipe.Thickness"),
        method = c("CM", "VM", "EJD", "GQ", "SPT"))
int_cor(mushroom.int, var_name1 = "Pileus.Cap.Width",
        var_name2 = "Stipe.Length", method = "CM")

Histogram-valued data (`hist_*`)

Compute mean, variance, covariance, and correlation for histogram-valued data with methods BG and L2W (cov/cor also support BD, B).

library(HistDAWass)

hist_mean(HistDAWass::BLOOD, var_name = "Cholesterol", method = "BG")
hist_var(HistDAWass::BLOOD, var_name = "Cholesterol", method = "L2W")

hist_cov(HistDAWass::BLOOD, var_name1 = "Cholesterol",
         var_name2 = "Hemoglobin", method = "BD")
hist_cor(HistDAWass::BLOOD, var_name1 = "Cholesterol",
         var_name2 = "Hemoglobin", method = "BG")

Data Format Conversion

Interval format conversions

Function	Description
`int_detect_format`	Detect the format of an interval-valued dataset
`int_convert_format`	Convert between interval formats
`int_list_conversions`	List all available format conversions
`to_all_interval_formats`	Convert intervals to all supported formats at once

Other conversion functions

Function	Description
`RSDA_format`	Convert conventional data to RSDA format
`set_variable_format`	One-hot encode set variables for RSDA format
`aggregate_to_symbolic`	Convert traditional data to symbolic data format

Interval Geometry

Function	Description
`int_width`	Width of each interval
`int_radius`	Radius of each interval
`int_center`	Center point of each interval
`int_midrange`	Half-range of each interval
`int_overlap`	Overlap measure between two interval variables
`int_containment`	Check if one interval contains another

Interval Position and Scale

Function	Description
`int_median`	Median of interval data
`int_quantile`	Quantiles of interval data
`int_range`	Range of interval data
`int_iqr`	Interquartile range
`int_mad`	Median absolute deviation
`int_mode`	Mode of interval data

Interval Shape

Function	Description
`int_skewness`	Skewness of interval data
`int_kurtosis`	Kurtosis of interval data
`int_symmetry`	Symmetry coefficient
`int_tailedness`	Tailedness measure

Interval Distance and Similarity

Function	Description
`int_dist`	Distance measures (GD, IY, L1, L2, CB, HD, EHD, WD, etc.)
`int_jaccard`	Jaccard similarity coefficient
`int_dice`	Dice similarity coefficient
`int_cosine`	Cosine similarity
`int_overlap_coefficient`	Overlap coefficient
`int_tanimoto`	Tanimoto coefficient
`int_similarity_matrix`	Pairwise similarity matrix

Interval Robust Statistics

Function	Description
`int_trimmed_mean`	Trimmed mean
`int_winsorized_mean`	Winsorized mean
`int_trimmed_var`	Trimmed variance
`int_winsorized_var`	Winsorized variance

Interval Uncertainty and Variability

Function	Description
`int_entropy`	Shannon entropy
`int_cv`	Coefficient of variation
`int_dispersion`	Dispersion index
`int_imprecision`	Imprecision based on interval width
`int_granularity`	Variability in interval sizes
`int_uniformity`	Uniformity of interval widths
`int_information_content`	Normalized entropy

Utilities

Function	Description
`clean_colnames`	Clean column names of a data frame
`read_symbolic_csv`	Read symbolic data from CSV file
`write_symbolic_csv`	Write symbolic data to CSV file
`search_data`	Search available datasets by keyword or type
`aggregate_to_symbolic`	Convert traditional data to symbolic data format

Datasets

The package includes 114 built-in datasets for symbolic data analysis:

Interval-valued datasets (53 datasets, `.int`)

abalone.int, acid_rain.int, age_cholesterol_weight.int, baseball.int, bats.int, blood_pressure.int, car.int, car_models.int, cardiological.int, cars.int, china_temp.int, china_temp_monthly.int, credit_card.int, ecoli_routes.int, employment.int, finance.int, freshwater_fish.int, fungi.int, genome_abundances.int, hdi_gender.int, horses.int, iris.int, judge1.int, judge2.int, judge3.int, lackinfo.int, lisbon_air_quality.int, loans_by_purpose.int, loans_by_risk.int, loans_by_risk_quantile.int, lynne1.int, mushroom.int, nycflights.int, ohtemp.int, oils.int, polish_voivodships.int, profession.int, prostate.int, soccer_bivar.int, synthetic_clusters.int, teams.int, temperature_city.int, tennis.int, trivial_intervals.int, uscrime.int, utsnow.int, veterinary.int, video1.int, video2.int, video3.int, water_flow.int, wine.int, world_cup.int

Histogram-valued datasets (25 datasets, `.hist`)

age_pyramids.hist, airline_flights.hist, bird_color_taxonomy.hist, blood.hist, china_climate_month.hist, china_climate_season.hist, cholesterol.hist, county_income_gender.hist, cover_types.hist, exchange_rate_returns.hist, flights_detail.hist, french_agriculture.hist, glucose.hist, hardwood.hist, hematocrit.hist, hematocrit_hemoglobin.hist, hemoglobin.hist, hierarchy.hist, hospital.hist, iris_species.hist, lung_cancer.hist, ozone.hist, simulated.hist, state_income.hist, weight_age.hist

Mixed symbolic datasets (11 datasets, `.mix`)

bird.mix, bird_species.mix, bird_species_extended.mix, census.mix, environment.mix, health_insurance.mix, joggers.mix, mtcars.mix, mushroom_fuzzy.mix, polish_cars.mix, town_services.mix

Interval time series datasets (9 datasets, `.its`)

crude_oil_wti.its, djia.its, euro_usd.its, ibovespa.its, irish_wind.its, merval.its, petrobras.its, shanghai_stock.its, sp500.its

Modal-valued datasets (7 datasets, `.modal`)

airline_flights2.modal, crime.modal, crime2.modal, fuel_consumption.modal, health_insurance2.modal, occupations.modal, occupations2.modal

Distribution-valued datasets (3 datasets, `.distr`)

energy_consumption.distr, energy_usage.distr, household_characteristics.distr

iGAP format datasets (2 datasets, `.iGAP`)

abalone.iGAP, face.iGAP

Other datasets

bank_rates, hierarchy, mushroom.int.mm

Vignettes

Introduction to dataSDA

Dependencies

R (>= 4.0.0)
RSDA, HistDAWass, dplyr, tidyr, magrittr, methods

Authors

Po-Wei Chen (Author), Chun-houh Chen (Author)
Han-Ming Wu (Creator, Maintainer) - wuhm@g.nccu.edu.tw

License

GPL (>= 2)