Search Configuration
Configure search settings for sanctions list
What is Search Configuration?
Search Configuration is a feature in Ambriel's AML solution that allows you to customize how sanctions lists are searched. You can configure the search settings to better suit your business needs.
What are the Configurations?
The following are the configurations you can make:
Frequency
The frequency determines how often the sanctions list is updated. The options are:
- Daily
- Weekly
- Bi-Weekly
- Quarterly
- Monthly
- Yearly
Enable Sanctions
Enable sanctions determines whether the sanctions list is enabled or not. If enabled, the sanctions list will be searched when you search for an entity.
Enable PEP
Enable PEP determines whether the PEP list is enabled or not. If enabled, the PEP list will be searched when you search for an entity.
Enable Crime
Enable crime determines whether the crime list is enabled or not. If enabled, the crime list will be searched when you search for an entity.
Enable Regulatory
Enable regulatory determines whether the regulatory list is enabled or not. If enabled, the regulatory list will be searched when you search for an entity.
Search Type
The search type determines how the search is performed. The options are:
- Exact
- Fuzzy
Score Threshold
The score_threshold determines the minimum similarity score required for a document to be considered a match. The score ranges from 0.0 (no similarity) to 1.0 (perfect match). Only results with a score equal to or above the specified threshold are returned.
| Score Threshold | Search Query | Match | Result |
|---|---|---|---|
| 0.4 | sarah is a developer | sarah developer | |
| 0.6 | sarah is a developer | sarah developer | |
| 0.8 | sarah is a developer | sarah developer | |
| 0.7 | steve jobs biography | steve bio | |
| 0.9 | steve jobs biography | steve bio | |
| 0.5 | full stack engineer | full engineer | |
| 0.8 | full stack engineer | full engineer |
Note
A lower score threshold allows more flexible (but potentially less accurate) matches. A higher score threshold enforces stricter matching.
Fuzzy Search Settings
The following are the fuzzy search settings you can make:
Edit Distance Thresholds
These thresholds control when typo-tolerant matching kicks in, based on the length of each token in the query.
min_len_1typo
A token must be at least this long to be allowed 1 typo. Default: 4
| min_len_1typo | Token | Match Token | Token Length | Allowed Typos | Match? |
|---|---|---|---|---|---|
| 3 | john | jon | 4 | 1 | Yes – meets length |
| 3 | jon | john | 3 | 1 | Yes – meets length |
| 5 | dave | david | 4 | 0 | No – token too short for typos |
| 5 | david | dave | 5 | 1 | Yes – meets threshold |
| 4 | mike | miek | 4 | 1 | Yes – meets threshold |
| 4 | amy | ami | 3 | 0 | No – too short for typo matching |
Edit distance threshold for two typos
min_len_2typos
A token must be at least this long to be allowed 2 typos. Default: 8
| min_len_2typos | Token/Query | Match | Longest Token Length | Allowed Typos | Match? |
|---|---|---|---|---|---|
| 5 | steve | steev | 5 | 1 | Yes – within 1 typo |
| 5 | steev | steve | 5 | 1 | Yes – within 1 typo |
| 7 | sarah is a developer | sarah is a dev | developer → 9 | ≤2 | Yes – allowed (token length > 7) |
| 7 | sarah is a dev | sarah is a developer | developer → 9 | ≤2 | Yes – allowed |
| 6 | daniel | danyel | 6 | 2 | Yes – allowed |
| 6 | danyel | daniel | 6 | 2 | Yes – allowed |
| 6 | ben | bann | 3 | 0 | No – too short for typos |
Num Typos
The num typos determines the maximum number of typos allowed in a search query.
| Num Typos | Search Query | Match | Result |
|---|---|---|---|
| 1 | steve | steev | |
| 1 | steev | steve | |
| 2 | sarah is a developer | sarah is a dev | |
| 2 | sarah is a dev | sarah is a developer |
Important
Typos are allowed per token, not across the entire phrase.
| Num Typos | Search Query | Match | Result |
|---|---|---|---|
| 2 | michael | mykael | |
| 1 | michael | mykael |
Drop Tokens Threshold
drop_tokens_threshold determines the minimum number of results needed before Search engine will begin dropping tokens (words) from the query. This parameter accepts only integer values.
0: Disables token dropping (behaves like a strict AND search).
≥1: If the search returns fewer than this number of results, Search engine will iteratively remove tokens (starting with those least likely to match) until enough results are found or only one token remains.
| Drop Tokens Threshold | Search Query | Behavior & Match |
|---|---|---|
| 0 | sarah is a developer | No token dropping – exact presence of all tokens required for a match. |
| 1 (default) | sarah is a dev | If fewer than 1 result, starts dropping “a”; if still too few, drops “is”, etc. |
| 2 | steve is a dev today | If fewer than 2 results, can drop “today” first, then “dev”, until ≥2 results. |
| 3 | java spring boot microservices | If < 3 results, drops least matching token (e.g., “microservices”), retry search. |
Use Case
Useful when search queries contain noise words (like "a", "is", "the", etc.), allowing better match scoring without requiring exact matches.
Tip
Adjusting drop_tokens_threshold lets you control how relaxed or strict the search should be regarding missing words.
| Use Case | Value |
|---|---|
| Strict exact-ish match | 0.8 or 0.9 |
| Allow for common noise words (is, the, a) | 0.5 to 0.6 |
| Require all tokens to match | 0 |
Typo Tokens Threshold
The typo_tokens_threshold parameter controls how many tokens in the search query are allowed to have typos (i.e., fuzzy matching). The remaining tokens must match exactly.
It does not limit the number of typos per token — that’s controlled by num_typos.
It does limit how many tokens in the query are eligible for typo tolerance.
Search Query: "Bernadette Hamill Johnson"
→ Tokens: ["Bernadette", "Hamill", "Johnson"]| Value (default=1) | Max Tokens Allowed with Typos | Match Behavior |
|---|---|---|
| 1 | 1 | Only 1 token can have typos, 2 tokens must match exactly |
| 2 | 2 | Up to 2 tokens can have typos, 1 must match exactly |
| 3 | 3 | All tokens can have typos (loosest match) |
Use Case Recommendations
| Use Case | typo_tokens_threshold Value |
|---|---|
| Strict exact-ish match | 1 |
| Allow for common noise words (is, the, a) | 2 |
| Loose fuzzy matching (name search) | 3 |
Remarks
Attention!
Loose fuzzy matching (name search) and less score threshold is not recommended. It can lead to unexpected/false-positive results. Be careful with this setting and use it with caution. Remember to always double-check the search results.
How to Configure Search Settings?
Login to your Ambriel dashboard.
Go to the Settings menu and select AML Configuration
Configure the search settings as desired.