Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.
Querykeyer
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In development
General information
Theme
Begin date
Contact information
Responsible use
Goal and impact
The police use keywords to find information in systems. To get good search results, it is important that these words match the language used in the records. That language often deviates from normal Dutch, for instance due to abbreviations and police terms.
The Querytoetser helps with this. Based on entered search words, the tool shows other matching words. In doing so, it takes into account the language used within the police.
The user can enter a word, e.g. the word 'firearm'. The application then shows other matching words, such as 'handgun', 'vuwa' and 'firearm-like object'.
By then using these words in a search query, employees can search more completely and find the right information faster.
Considerations
If a search is incomplete, important registrations may not be found. This application helps to make a search more complete. As a result, relevant information is more likely to be found.
Human intervention
The user decides which words are used in the search. As a result, the model does not affect the final search the employee makes.
Risk management
It is up to the user to be legally and ethically aware when creating a query.
A word must appear at least 15 times in the police system, otherwise it is too unique to be included in the model.
Names of offenders are removed from texts.
Operations
Data
Free text data from police systems (dates from 01-10-2024 to 01-10-2025)
Technical design
The application uses a model that compares words based on usage. The model learns which words often appear together in texts.
To do so, it uses a simple two-layer neural network. The model predicts words based on surrounding words. It thus learns which words are similar in content. Words with similar meanings therefore appear close together in the model.
Similar algorithm descriptions
- When conducting elections, data of polling station members and volunteers is maintained in the application Polling Station Manager. Two types of algorithms are deployed, namely 1. location algorithm to show volunteers the voting locations in their own area and 2. experience algorithm that uses the experience of volunteers to suggest appropriate positions and trainingLast change on 31st of October 2025, at 13:28 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Uthiek
- Status
- In use
- The search algorithm used by the BSN management facility to search for individuals based on identifying data.Last change on 25th of September 2024, at 13:20 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- DPIA
- Status
- In use