Predictive Hotel Id matching API
beta program. Free access
- No data stored
- No database needed
- No third parties dependencies
- Transparent funnel
How it works
An anonymous standalone.
Our ubiquitous predictive logic can receive any data type, across X number of arguments.
Predictive mapping works as a standalone, down to it s core. The same ubiquitous algorithm is applied for any data.
By using a NLP dual trust scoring model, it can match ids on various scales and refine his trust judgment. see below
check demo sandbox tab to see how it works.
Standardized standalone
Arguments used for matching:
name, adr, city, cntry, zip, iso2, desti, other
wysiwyg approach:
What you send is what you get
You can pass any number of arguments as you want, independently per line, for both the input and versus.
Using at least a name or an address argument is best practice. more arguments means a more precise result.
Transparent funnel. No suppliers or ids needed. results are returned as they are sent.
Dual trust authority
2 matching % results are returned:
The trust % is the match trust % based on each argument weight and score.
This is an holistic analysis across all arguments, precisely reflecting tiny differences.
case study: You have a resort on an island, same name, same address, with 3 aisles (north, south, east), having different amenities or star rating. You do not want to consider them as duplicates, despite same address, name, gps...
The confi % is the duplicate confidence ratio based additional correlative algorithms, amplifying the trust ratio curve (positively or negatively).
This allows to find out of the bat more direct matches and duplicates.
case study: You want to sort out duplicates, match ids, qualify sub groups, chains, rentals. Additional correlative algorithms are applied, amplifying the trust ratio curve (positively or negatively). Can be custom per user. very lightly amplified in the demo.
Precision for unique & refined view to match duplicates.
APIs scoring NLP model
In Natural language Processing (NLP) a weight defines the importance of an argument. Those weights are used for scoring. The scoring defines the % of dissimilarities.In addition to the usual arguments, you have at your disposal 2 more (desti & other) for your own purposes (criteria, filters...) Use them in accordance to the weight you want to allocate them. Check below weight table
| argument | description | weight | Optional |
|---|---|---|---|
| hid | your reference | ||
| name | The hotel name | heavier | |
| adr | The hotel address | heavier | |
| city | The city name | average | |
| cntry | The country name | light | |
| zip | The postal code | heavy | |
| iso2 | The iso 2 country code | average | |
| desti | Destination or alike, other | light | |
| other | Your own, can be anything | heavy | |
Use Demoversus credential to try it free.
Copy / paste your own json into the sandbox string box, or use one of our 3 demo files, and click send to predictive API button.
Instructions
Look for any trust over 90% as a direct match
In 95% of cases, highest % returned is the match or no match exist
Query
const request = $.ajax({type: "POST",datatype: "text",url: `https://travelalliance.com/MAP/versus`,data: queryJSON});
Response
request.done(response => {responseJSON});
query
Note in this example, the versus one is without name.
response
Note the 2 returned matching % arguments, trust and confi.