Element Scoring

Scoring configuration is done with the help of pseudo inputs that are to be used in any of the expressions that the Scoring provides. The pseudo inputs are as follows:

ia - Input address
pa - Proposal address
expl - Explanation address
const - Constants for explanation. This represents a constant record with the fixed values: missing, found, approx, congruity, not_defined and other. Its sole purpose is to compare with values in the expl pseudo input
e - Count of different characters during approximate searching
z - Count of missing characters during truncation of a word. The truncation is triggered either by a dot character or a space or end of input. If it was triggered by a dot, this dot is not considered as unexplained text.
d - Count of characters with a missing diacritic
c - Resemblance of the input component to the proposal component - currently usable only for Czech ZIP codes (PS). It is defined as the position (counted from the right) of the first different character from the left. In other words, the length of the suffix of the input value, where the first character differs from the character at the same position in the proposal value (other characters of the suffix might or might not be same) and at the same time the prefix is the same for both.
v - Input values of the congruent components. Each component defined as congruent has its column and is to be filled with an input value that is detected as congruent to the reference data one. If the input value is not detected as congruent, no column exists for it. Note that ia.psc might be the same as this value but not necessarily, since there might exist situations with a congruent value on input which is not present in the dictionary (e.g., 99999).
x - Count of letters occurring in the input text in addition to the dictionary value. An example of this case is the street B. Nemcove, which might be expanded to Bozeny Nemcove, and in such case the value of the input for the street is 5. The expansion might occur more than once, such as in: Jaromerice n. R. -> Jaromerice nad Rokytnou, where the value is 9. This expansion is available only for dictionary values that contain dot(s).

Each pseudo-input except the const input includes address components, which are declared in the reference data configuration. Moreover, there is a component called unexpl in the pa input, whose value returns unexplained text.

Name	Type	Required	Description
When	String	Yes	Determines when score and explanation values are written into the resulting score. Default value: None.
Score	String	Yes	Integer expression whose result is added to the score column after accomplishing the when condition. Caution: if the scoring for the input address evaluates to null, the overall scoring is also null. Default value: None.
Explanation	String	Yes	String expression whose result is appended to the explanation column after accomplishing the when condition. Unlike the score property, this expression may evaluate to null in which cases the "null" string is appended to the result to emphasize that such scoring was effective. Default value: None.

Top of page

Example: Example

<Scoring>
  <explanation>'ulice_Novakova'</explanation>
  <score>666</score>
  <when>ia.ulice == 'Novakova'</when>
</Scoring>

If there is street called ‘Novakovatm) in the input it will add the score 666 and append the text 'ulice_Novakova'.

<Scoring>
  <explanation>'prop_obec_Kocourkov'</explanation>
  <score>12347</score>
  <when>pa.obec == 'Kocourkov'</when>
</Scoring>

If there is city called ‘kocourkovtm) in the proposal it will add the score 12347 and append the text 'prop_obec_Kocourkov'

<Scoring>
  <explanation>'po_MISSING'</explanation>
  <score>10</score>
  <when>expl.po == const.missing</when>
</Scoring>

If the post name is missing it will add the score 10 and append the text 'po_MISSING

<Scoring>
  <explanation>'ulice_APPROX_E_' + toString(e.ulice)</explanation>
  <score>10</score>
  <when>e.ulice > 0</when>
</Scoring>

If the street is found with an approximate name it will add the score 10 and append the text 'ulice_APPROX_E_(count of different chars)

<Scoring>
  <explanation>'ulice_APPROX_Z_' + toString(z.ulice)</explanation>
  <score>10</score>
  <when>z.ulice > 0</when>
</Scoring>

If the street is found with a truncated name it will add the score 10 and append the text 'ulice_APPROX_Z_(count of missing chars)

<Scoring>
  <explanation>'ulice_APPROX_D_' + toString(d.ulice)</explanation>
  <score>10</score>
  <when>d.ulice > 0</when>
</Scoring>

If the street is found with missing diacritics it will add the score 10 and append the text 'ulice_APPROX_D_(count of chars with missing diacritic)

<Scoring>
  <explanation>'UNEXPL_' + toString(pa.unexpl)</explanation>
  <score>length(pa.unexpl) * 10</score>
  <when>length(pa.unexpl) > 0</when>
</Scoring>

If there are unexplained chars it will add the score 10 * the count of unexplained chars and append the text 'unexpl_(count of unexplained chars)

<Scoring>
  <explanation>'PSC_CHARS1'</explanation>
  <score>5</score>
  <when>c.psc == 1</when>
</Scoring>

If the ZIP differs on the last character it will add the score 5 and append the text 'ZIP_CHARS1'. Exemplary configuration for Dictionary Lookup Identifier:

<scoring>
  <scoring when="expl.ulice == const.approx AND z.ulice &gt; 0" score="10" explanation="'ulice_SHORT'"/>
  <scoring when="expl.obec == const.approx AND z.obec &gt; 0" score="10" explanation="'obec_SHORT'"/>
  <scoring when="expl.cobce == const.approx AND z.cobce &gt; 0" score="10" explanation="'cobce_SHORT'"/>
  <scoring when="expl.posta == const.approx AND z.posta &gt; 0" score="10" explanation="'po_SHORT'"/>
  <scoring when="expl.okres == const.approx AND z.okres &gt; 0" score="10" explanation="'okres_SHORT'"/>
  <scoring when="expl.mcast == const.approx AND z.mcast &gt; 0" score="10" explanation="'mcast_SHORT'"/>
  <scoring when="expl.ulice == const.approx AND e.ulice &gt; 0" score="10*e.ulice" explanation="'ulice_APPROX'+toString(e.ulice)"/>
  <scoring when="expl.obec == const.approx AND e.obec &gt; 0" score="10*e.obec" explanation="'obec_APPROX'+toString(e.obec)"/>
  <scoring when="expl.cobce == const.approx AND e.cobce &gt; 0" score="10*e.cobce" explanation="'cobce_APPROX'+toString(e.cobce)"/>
  <scoring when="expl.posta == const.approx AND e.posta &gt; 0" score="10*e.posta" explanation="'po_APPROX'+toString(e.posta)"/>
  <scoring when="expl.okres == const.approx AND e.okres &gt; 0" score="10*e.okres" explanation="'okres_APPROX'+toString(e.okres)"/>
  <scoring when="expl.mcast == const.approx AND e.mcast &gt; 0" score="10*e.mcast" explanation="'mcast_APPROX'+toString(e.mcast)"/>
  <scoring when="expl.ulice == const.approx AND d.ulice &gt; 0" score="d.ulice" explanation="'ulice_DIA'+toString(d.ulice)"/>
  <scoring when="expl.obec == const.approx AND d.obec &gt; 0" score="d.obec" explanation="'obec_DIA'+toString(d.obec)"/>
  <scoring when="expl.cobce == const.approx AND d.cobce &gt; 0" score="d.cobce" explanation="'cobce_DIA'+toString(d.cobce)"/>
  <scoring when="expl.posta == const.approx AND d.posta &gt; 0" score="d.posta" explanation="'po_DIA'+toString(d.posta)"/>
  <scoring when="expl.okres == const.approx AND d.okres &gt; 0" score="d.okres" explanation="'okres_DIA'+toString(d.okres)"/>
  <scoring when="expl.mcast == const.approx AND d.mcast &gt; 0" score="d.mcast" explanation="'mcast_DIA'+toString(d.mcast)"/>
  <scoring when="expl.ulice == const.missing" score="100" explanation="'ulice_MISSING'"/>
  <scoring when="expl.cobce == const.missing AND pa.ulice == null" score="100" explanation="'cobce_missing'"/>
  <scoring when="expl.posta == const.missing AND expl.obec == const.found AND indexOf(pa.posta, pa.obec) == 0" score="0" explanation="'po_City_INCLUDED'"/>
  <scoring when="expl.posta == const.missing AND expl.obec == const.found AND pa.posta != null AND pa.posta != pa.obec" score="10" explanation="'po_missing'"/>
  <scoring when="expl.obec == const.missing AND expl.posta == const.found AND indexOf(pa.obec, pa.posta) == 0" score="0" explanation="'obec_Post_INCLUDED'"/>
  <scoring when="expl.obec == const.missing AND expl.posta == const.found AND pa.obec != null AND pa.obec != pa.posta" score="10" explanation="'city_missing'"/>
  <scoring when="expl.obec == const.missing AND expl.posta == const.missing" score="100" explanation="'City&amp;Post_missing'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 1" score="5" explanation="'psc_CHARS1'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 2" score="10" explanation="'psc_CHARS2'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 3 AND substr(v.psc, 0, 1) == '1'" score="10" explanation="'psc_CHARS3'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 3 AND substr(v.psc, 0, 1) != '1'" score="100" explanation="'psc_CHARS3'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 4" score="200" explanation="'psc_CHARS4'"/>
  <scoring when="expl.psc == const.congruity AND c.psc == 5" score="400" explanation="'psc_CHARS5'"/>
  <scoring when="length(pa.unexpl) &gt; 3" score="20*length(pa.unexpl)" explanation="'UNEXPLAINED_CHARS'"/>
  <scoring when="length(pa.unexpl) &gt; 0 AND length(pa.unexpl) &lt;= 3" score="10*length(pa.unexpl)" explanation="'UNEXPLAINED_CHARS'"/>
</scoring>

iWay Software