Matome

A program that generates linguistic summarizations of a body performance dataset¹.

Both, single-subject and multiple-subject summaries can be generated of the forms from I to IV. The quality of the summaries is determined by the $T_1$ - $T_{11}$ measures².

Quick start

• Download the dataset from https://www.kaggle.com/datasets/kukuroo3/body-performance-data and put it in the project's top-level directory as bodyPerformance.csv;
• Execute the following commands:

  $ python preprocess_dataset.py
  $ mvn javafx:run

Running in CLI mode

Warning

Execution of the program in CLI mode with the default config.json may take significant amount of time. You may want to reduce the number of quantifiers and qualifiers first in the config.json file.

In order to run the program without GUI, the dataset needs to be downloaded and preprocessed as in the previous step, and the following commands need to be executed:

$ mvn compile assembly:single
$ java -jar target/matome-<VERSION>-SNAPSHOT-jar-with-dependencies.jar [OPTIONS]

The <VERSION> part in the last command needs to be replaced by a specific version, e.g. 1.0-SNAPSHOT.

The available OPTIONS can be obtained by running the last command with the --help option.

Sample summaries

The following table shows sample summaries generated with Matome:

Summary	T	Form
About 2000 people are tall	0.781	I
About 1/4 people being tall have very strong grip force	0.914	II
About 3/4 females compared to males are short and have normal grip force	0.916	I
About half females, compared to males being young, are medium height	0.981	II
Almost all females being short, compared to males, are underweight (BMI)	0.995	III
More males than females have moderate broad jump	0.547	IV

Attribures

Below is a list of attributes used to generate summaries. Most of them come from the body performance dataset, but some are generated by the preprocess_dataset.py script.

• age, expressed by an integer between 21 and 64, inclusivly.
• gender, expressed by one of two values: M, representing the male gender and F, representing the female gender.
• height in centimeters, expressed by a real number to one decimal place between 125.0 and 194.0.
• weight in kilograms, expressed by a real number to one decimal place between 26.3 and 138.0.
• BMI (Body Mas Index) in kilograms per square meter, expressed by a real number to one decimal place between 0.0 and 90.0, generated based on weight and height from the following formula:
  $$BMI=\frac{weight}{(0.01\cdot height)^2}$$
• body fat (percentage), expressed by a real number to one decimal place between 3.0 and 78.4.
• modified body fat, generated based on the body fat and gender attributes. If the value of the gender attribute of a record is M, the value of this attribute is the same as the value of the body fat attribute. Otherwise, the value of this attribute is the value of the body fat attribute minus 8.2. This is because women have an average of 8.2 more body fat percentage than men³.
• diastolic, i.e. blood pressure in the diastolic phase of the cardiac cycle, expressed by an integer between 0 and 156.
• systolic, i.e. blood pressure in the systolic phase of the cardiac cycle, expressed by an integer between 0 and 201.
• grip force in kilograms measured with a handheld dynanometer, expressed as a real number to one decimal place between 0.0 and 70.5.
• sit and bend forward, i.e. the longest distance in centimeters that the tips of the fingers travel horizontally during forward bending in the sitting position, expressed by a real number to one decimal place between -25.0 and 42.0. A negative value of this attribute means that the person was not able to assume an upright sitting position (i.e., at 90°) and the given value is the path needed to assume this position.
• sit-ups count, i.e. the number of sit-ups performed in 2 minutes, expressed by an integer between 0 and 80.
• broad jump length in centimeters, expressed by an integer between 0 and 303.

Default configuration

Linguistic variables

• age
  Domain: [20, 70]
  $\begin{align} \mathrm{age}_{\mathrm{YOUNG}}(x) &= \begin{cases} 1 & \text{ for } x < 30 \\ \frac{50-x}{20} & \text{ for } x \in [30, 50) \end{cases} \\ \mathrm{age}_{\mathrm{AVERAGE}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-36.8}{2}\right)^2\right) \\ \mathrm{age}_{\mathrm{MIDDLE-AGED}}(x) &= \begin{cases} \frac{x-35}{10} & \text{ for } x \in [35, 45) \\ 1 & \text{ for } x \in [45, 55) \\ \frac{65-x}{10} & \text{ for } x \in [55, 65) \\ \end{cases} \\ \mathrm{age}_{\mathrm{OLD}}(x) &= \begin{cases} \frac{x-50}{20} & \text{ for } x \in [50, 70) \\ 1 & \text{ for } x \ge 70 \\ \end{cases} \end{align}$
• BMI
  Domain: [0, 90]
  $\begin{align} \mathrm{BMI}_{\mathrm{UNDERWEIGHT}}(x) &= \begin{cases} 1 & \text{ for } x < 16 \\ \frac{21-x}{5} & \text{ for } x \in [16, 21) \\ \end{cases} \\ \mathrm{BMI}_{\mathrm{NORMAL-WEIGHT}}(x) &= \begin{cases} \frac{x-16.5}{4} & \text{ for } x \in [16.5, 20.5) \\ 1 & \text{ for } x \in [20.5, 23) \\ \frac{27-x}{4} & \text{ for } x \in [23, 27) \\ \end{cases} \\ \mathrm{BMI}_{\mathrm{OVERWEIGHT}}(x) &= \begin{cases} \frac{x-23}{4} & \text{ for } x \in [23, 27) \\ 1 & \text{ for } x \in [27, 28) \\ \frac{32-x}{4} & \text{ for } x \in [28, 32) \\ \end{cases} \\ \mathrm{BMI}_{\mathrm{OBESE}}(x) &= \begin{cases} \frac{x-28}{4} & \text{ for } x \in [28, 32) \\ 1 & \text{ for } x \ge 32 \\ \end{cases} \end{align}$
• modified body fat
  Domain: [0, 80]
  $\begin{align} \textrm{modified-body-fat}_{\mathrm{LOW}}(x) &= \begin{cases} 1 & \text{ for } x < 4 \\ \frac{20-x}{16} & \text{ for } x \in [4, 20) \\ \end{cases} \\ \textrm{modified-body-fat}_{\mathrm{IDEAL}}(x) &= \begin{cases} \frac{x+2}{16} & \text{ for } x \in [-2, 14) \\ 1 & \text{ for } x \in [14, 16) \\ \frac{26-x}{10} & \text{ for } x \in [16, 26) \\ \end{cases} \\ \textrm{modified-body-fat}_{\mathrm{AVERAGE}}(x) &= \begin{cases} \frac{x-12}{8} & \text{ for } x \in [12, 20) \\ 1 & \text{ for } x \in [20, 22) \\ \frac{34-x}{12} & \text{ for } x \in [22, 34) \\ \end{cases} \\ \textrm{modified-body-fat}_{\mathrm{HIGH}}(x) &= \begin{cases} \frac{x-16}{8} & \text{ for } x \in [16, 24) \\ 1 & \text{ for } x \ge 24 \\ \end{cases} \end{align}$
• broad jump
  Domain: [0, 310]
  $\begin{align} \textrm{broad-jump}_{\textrm{SHORT}}(x) &= \begin{cases} 1 & \text{ for } x < 125 \\ \frac{175-x}{50} & \text{ for } x \in [125, 175) \\ \end{cases} \\ \textrm{borad-jump}_{\textrm{MODERATE}}(x) &= \begin{cases} \frac{x-125}{50} & \text{ for } x \in [125, 175) \\ 1 & \text{ for } x \in [175, 225) \\ \frac{275-x}{50} & \text{ for } x \in [225, 275) \\ \end{cases} \\ \textrm{broad-jump}_{\textrm{LONG}}(x) &= \begin{cases} \frac{x-225}{85} & \text{ for } x \in [225, 310) \\ 1 & \text{ for } x \ge 310 \\ \end{cases} \end{align}$
• diastolic
  Domain: [0, 160]
  $\begin{align} \mathrm{diastolic}_{\mathrm{HYPOTENSION}}(x) &= \begin{cases} 1 & \text{ for } x < 50 \\ \frac{70-x}{20} & \text{ for } x \in [50, 70) \\ \end{cases} \\ \mathrm{diastolic}_{\mathrm{NORMAL}}(x) &= \begin{cases} \frac{x-50}{20} & \text{ for } x \in [50, 70) \\ 1 & \text{ for } x \in [70, 80) \\ \frac{100-x}{20} & \text{ for } x \in [80, 100) \\ \end{cases} \\ \mathrm{diastolic}_{\mathrm{HYPERTENSION}}(x) &= \begin{cases} \frac{x-80}{20} & \text{ for } x \in [80, 100) \\ 1 & \text{ for } x \ge 100 \\ \end{cases} \end{align}$
• systolic
  Domain: [0, 210]
  $\begin{align} \mathrm{systolic}_{\mathrm{HYPOTENSION}}(x) &= \begin{cases} 1 & \text{ for } x < 70 \\ \frac{110-x}{40} & \text{ for } x \in [70, 110) \\ \end{cases} \\ \mathrm{systolic}_{\mathrm{NORMAL}}(x) &= \begin{cases} \frac{x-70}{40} & \text{ for } x \in [70, 110) \\ 1 & \text{ for } x \in [110, 120) \\ \frac{160-x}{40} & \text{ for } x \in [120, 160) \\ \end{cases} \\ \mathrm{systolic}_{\mathrm{HYPERTENSION}}(x) &= \begin{cases} \frac{x-120}{40} & \text{ for } x \in [120, 160) \\ 1 & \text{ for } x \ge 160 \\ \end{cases} \end{align}$
• grip force
  Domain: [0, 80]
  $\begin{align} \textrm{grip-force}_{\textrm{VERY-WEAK}}(x) &= \begin{cases} 1 & \text{ for } x < 0 \\ \frac{16-x}{16} & \text{ for } x \in [0, 16) \\ \end{cases} \\ \textrm{grip-force}_{\textrm{WEAK}}(x) &= \begin{cases} \frac{x-16}{4} & \text{ for } x \in [16, 20) \\ 1 & \text{ for } x \in [20, 30) \\ \frac{34-x}{4} & \text{ for } x \in [30, 34) \\ \end{cases} \\ \textrm{grip-force}_{\textrm{NORMAL}}(x) &= \begin{cases} \frac{x-30}{4} & \text{ for } x \in [30, 34) \\ \frac{38-x}{4} & \text{ for } x \in [34, 38) \\ \end{cases} \\ \textrm{grip-force}_{\textrm{STRONG}}(x) &= \begin{cases} \frac{x-34}{4} & \text{ for } x \in [34, 38) \\ 1 & \text{ for } x \in [38, 50) \\ \frac{54-x}{4} & \text{ for } x \in [50, 54) \\ \end{cases} \\ \textrm{grip-force}_{\textrm{VERY-STRONG}}(x) &= \begin{cases} \frac{x-50}{4} & \text{ for } x \in [50, 54) \\ 1 & \text{ for } x \ge 54 \\ \end{cases} \end{align}$
• height
  Domain: [120, 200]
  $\begin{align} \mathrm{height}_{\mathrm{SHORT}}(x) &= \begin{cases} 1 & \text{ for } x < 150 \\ \frac{170-x}{20} & \text{ for } x \in [150, 170) \\ \end{cases} \\ \mathrm{height}_{\mathrm{MEDIUM-HEIGHT}}(x) &= \begin{cases} \frac{x-150}{20} & \text{ for } x \in [150, 170) \\ \frac{190-x}{20} & \text{ for } x \in [170, 190) \\ \end{cases} \\ \mathrm{height}_{\mathrm{TALL}}(x) &= \begin{cases} \frac{x-170}{190-170} & \text{ for } x \in [170, 190) \\ 1 & \text{ for } x \ge 190 \\ \end{cases} \end{align}$
• sit and bend forward
  Domain: [-25, 220]
  $\begin{align} \textrm{sit-and-bend-forward}_{\textrm{MINIMAL}}(x) &= \begin{cases} 1 & \text{ for } x < 5 \\ \frac{15-x}{10} & \text{ for } x \in [5, 15) \\ \end{cases} \\ \textrm{sit-and-bend-forward}_{\textrm{MODERATE}}(x) &= \begin{cases} \frac{x-5}{10} & \text{ for } x \in [5, 15) \\ \frac{25-x}{10} & \text{ for } x \in [15, 25) \\ \end{cases} \\ \textrm{sit-and-bend-forward}_{\textrm{ADVANCED}}(x) &= \begin{cases} \frac{x-15}{10} & \text{ for } x \in [15, 25) \\ 1 & \text{ for } x \in [25, 30) \\ \frac{40-x}{10} & \text{ for } x \in [30, 40) \\ \end{cases} \\ \textrm{sit-and-bend-forward}_{\textrm{SUPERHUMAN}}(x) &= \begin{cases} \frac{x-30}{50} & \text{ for } x \in [30, 80) \\ 1 & \text{ for } x \ge 80 \\ \end{cases} \end{align}$
• sit-ups count
  Domain: [0, 80]
  $\begin{align} \textrm{sit-ups-count}_{\textrm{BEGINNER}}(x) &= \begin{cases} 1 & \text{ for } x < 10 \\ \frac{30-x}{10} & \text{ for } x \in [10, 30) \\ \end{cases} \\ \textrm{sit-ups-count}_{\textrm{INTERMEDIATE}}(x) &= \begin{cases} \frac{x-10}{20} & \text{ for } x \in [10, 30) \\ \frac{50-x}{20} & \text{ for } x \in [30, 50) \\ \end{cases} \\ \textrm{sit-ups-count}_{\textrm{ADVANCED}}(x) &= \begin{cases} \frac{x-30}{50} & \text{ for } x \in [30, 80) \\ 1 & \text{ for } x \ge 80 \\ \end{cases} \end{align}$

Quantifiers

• Relative quantifiers
  $\begin{align} \mu_{\textrm{ALMOST-NONE}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-0}{0.1}\right)^2\right) \\ \mu_{\textrm{ABOUT 1/4}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-0.25}{0.1}\right)^2\right) \\ \mu_{\textrm{ABOUT-HALF}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-0.5}{0.1}\right)^2\right) \\ \mu_{\textrm{ABOUT 3/4}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-0.75}{0.1}\right)^2\right) \\ \mu_{\textrm{ALMOST-ALL}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-1}{0.1}\right)^2\right) \end{align}$
• Absolute quantifiers
  $\begin{align} \mu_{\textrm{LESS-THAN-1000}}(x) &= \begin{cases} 1 & \text{ for } x < 500 \\ \frac{1000-x}{500} & \text{ for } x \in [500, 1000) \\ \end{cases} \\ \mu_{\textrm{ABOUT-2000}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-2000}{2000}\right)^2\right) \\ \mu_{\textrm{BETWEEN-4000-AND-6000}}(x) &= \begin{cases} \frac{x-4000}{500} & \text{ for } x \in [4000, 4500) \\ 1 & \text{ for } x \in [4500, 5500) \\ \frac{6000-x}{500} & \text{ for } x \in [5500, 6000) \\ \end{cases} \\ \mu_{\textrm{ABOUT-8000}}(x) &= \exp\left(-\frac{1}{2}\left(\frac{x-8000}{2000}\right)^2\right) \\ \mu_{\textrm{MORE-THAN-10000}}(x) &= \begin{cases} \frac{x-10000}{500} & \text{ for } x \in [10000, 10500) \\ 1 & \text{ for } x \ge 10500 \\ \end{cases} \end{align}$

LICENSE

MIT

Footnotes

1. Korea Sports Promotion Foundation, Body Performance Data, Kaggle. URL: https://www.kaggle.com/datasets/kukuroo3/body-performance-data. ↩

2. A. Niewiadomski, Methods for the Linguistic Summarization of Data: Applications of Fuzzy Sets and Their Extensions, Akademicka Oficyna Wydawnicza EXIT, Warszawa 2008. ↩

3. Cureton, K. J., Hensley, L. D., & Tiburzi, A. (1979). Body Fatness and Performance Differences between Men and Women. Research Quarterly. American Alliance for Health, Physical Education, Recreation and Dance. https://doi.org/10.1080/00345377.1979.10615619. ↩