Skip to main content

Table 3 Definitions of individual indicators

From: Identification of conclusive association entities in biomedical articles

Type Indicator Definition
(1) Frequency-based TF TF(e, a) = Number of times e appears in a
(2) Rareness-based IDF \( IDF(e)={Log}_2\frac{\mid A\mid +{1}^{\left[\mathrm{i}\right]}}{DF(e)+{1}^{\left[\mathrm{i}\mathrm{i}\right]}} \)
(3) Co-occurrence-based CoOcc \( CoOcc\left(e,a\right)=\sum \limits_{x\in a,x\ne e}\frac{{\left|{S}_{e\cap x}(a)\right|}^{\left[\mathrm{iii}\right]}}{{\left|{S}_e(a)\right|}^{\left[\mathrm{iv}\right]}} \)
(4) Concentration-based AvgTF \( AvgTF(e)=\frac{c{\left(e,C\right)}^{\left[\mathrm{v}\right]}}{DF(e)} \)
(5) Locality-based TITLE \( TITLE\left(e,a\right)=\left\{\begin{array}{c}1,\mathrm{if}\ \mathrm{e}\ \mathrm{a}\mathrm{ppears}\ \mathrm{in}\ \mathrm{title}\ \mathrm{of}\ \mathrm{a};\\ {}0,\mathrm{otherwise}.\end{array}\right. \)
AbstractX \( AbstractX\left(e,a\right)=\left\{\begin{array}{c}1,\mathrm{if}\ \mathrm{e}\ \mathrm{a}\mathrm{ppears}\ \mathrm{in}\ \mathrm{the}\ \mathrm{first}\ \mathrm{X}\ \mathrm{or}\ \mathrm{Last}\ \mathrm{X}\\ {}\ \mathrm{sentences}\ \mathrm{in}\ \mathrm{abstract}\ \mathrm{of}\ \mathrm{a};\\ {}0,\mathrm{otherwise}.\end{array}\right. \)
  1. [i] |A| = Number of articles in the collection of articles C;
  2. [ii]DF(e) = Number of articles (in C) mentioning e;
  3. [iii]Sex(a) = Set of sentences (in a) that e and x co-occur;
  4. [iv]Se(a) = Set of sentences (in a) mentioning e;
  5. [v]c(e,C) = Number of times e appears in articles in C