Show TOC

  Ranking

Use

The system uses this function to sort the search results in certain search scenarios according to their relevance in terms of the search criteria in a search query. The search criteria are weighted differently in the search.

You define the weighting of the search criteria beforehand in the respective search query. The system identifies entries that correspond to the search query and sorts these according to their weighting. In this way, you obtain a hit list in which the entries are specified with the respective ranking result.

Integration

This function is used in the standard SAP system for the search scenarios that the succession planner can use. The function is part of the search.

Prerequisites

To use the function, you must make the relevant settings in the Implementation Guide for SAP E-Recruiting .

Define Scales for Ranking and Reporting

In this activity, you define the scale that you want to use for weighting the search criteria. The values defined in the scale are used as follows when calculating the ranking:

  • The proficiency with the lowest numeric value for the ID is given the weighting 0 (minimum weighting)

  • The proficiency with the greatest numeric value for the ID is given the weighting 10000 (maximum weighting)

  • The weighting for all proficiencies inbetween is obtained from the numeric values for the IDs for the proficiencies. The following formula applies for calculating the weighting for an ID:

    Weighting for a particular ID = minimum weighting + [(numeric value of ID – numeric value of smallest ID)/(numeric value of greatest ID – numeric value of smallest ID)] * maximum weighting

The following examples illustrate how the values for the weighting that the system uses for calculating the ranking value are related to the numeric values of the IDs for the scale proficiencies.

Example 1

ID

Proficiency

Value for the weighting

1

Nice to have

0

2

Important

5000

3

Very important

10000

Example 2

ID

Proficiency

Value for the weighting

10

Nice to have

0

20

Important

5000

30

Very important

10000

Example 3

ID

Proficiency

Value for the weighting

5

Nice to have

0

10

Important

3333,3

20

Very important

10000

Assign Scales to Scale Types

In this activity, you assign the previously created scale to the Weighting for Search & Match scale type.

Set Up Search Applications

In this activity, you set up the search applications that are to output a hit list with ranking.

You make the following settings for the respective search applications:

  • Weighting indicator

    You set the indicator so that you can use weighted search criteria.

  • Ranking indicator

    This indicator enables the result of a ranking to be displayed in the hit list with the respective percentage.

  • Mandatory Values indicator

    You set this indicator if you only want entries in the hit list that fulfill search criteria previously identified as necessary search criteria using the weighting function.

  • Maximum Number of Hits

    You use this activity to define how many hits are to be displayed in the hit list.

  • Threshold

    You use this value to define by how much percent a hit must agree with the weighted search criteria for it to be displayed in the hit list.

Features

The ranking in the SAP E-Recruiting hit lists is based on internal calculations of the search engine TREX ( Text Retrieval and Information Extraction ).

TREX works in this way:

  • A search with multiple search criteria returns a list of pairs consisting of documents (hits) and ranking values.

  • The ranking value of a document (hit) indicates how well the document matches the search term. When calculating this value, the search engine considers the frequency of the search term and the number of documents found containing this term.

  • The value lists of the documents of the individual search criteria are linked using Boolean algebra so that the ranking values are calculated according to the operators.

Activities

The succession planner can influence the ranking result by weighting the search criteria differently. The succession planner chooses Weighting in the relevant search scenario and assigns weightings to the search criteria using a value scale.

Using the values resulting from the weighting and based on the number of documents found and the number of times a search item appears in the document, the system calculates the ranking values and sorts the hit list accordingly. The hits that best match the search criteria are displayed at the top of the list. The calculated ranking result for each hit is displayed.

Example

The succession planner wants to search in the Talent Pool for candidates who can speak English and are willing to travel. The succession planner rates the English language skill as very important, whereas the candidates’ willingness to travel is less important.

These conditions can be cataloged using the following search criteria, for example:

Search Criterion

Search Term

Weighting

Language Skills

English

Very important

Willingness to travel

Yes

Important

The succession planner performs the search as follows:

  1. On the search screen, the succession planner selects English as the search term for the criterion Language skills and Yes as the search term for the criterion Willingness to travel .

  2. Using the Weighting function, the succession planner applies the predefined value scale to weight the English language skills as Very important and the willingness to travel as Important .

  3. The succession planner then launches the search .

The system displays a hit list, with candidates at the top of the list who fulfill all search criteria and are ranked higher, followed by candidates who only fulfill one of the two criteria. Candidates who fulfill a criterion with higher weighting have a higher ranking result and are displayed higher up in the hit list. In this example, candidates who fulfill the English language skills criterion are displayed above candidates who fulfill the willingness to travel criterion. The ranking result calculated for each hit is displayed.