US-20260126972-A1 - DATA TRANSFORMATION SYSTEM

US20260126972A1US 20260126972 A1US20260126972 A1US 20260126972A1US-20260126972-A1

Abstract

A data transformation system is provided for extracting source data and transmitting it to a transformation processing unit. The transformation processing unit transforms the source data into standard format data through a mapping profile. By collecting format descriptions of the source data and data examples, in cooperation with documentation and a specification file of a standard format and a specification file of the mapping profile, a language model algorithm is used to create the mapping profile.

Inventors

Ta Wei Huang
Yu Ru Chen
Ying Ju Lin
YU TING CHEN
David Chen
Shan Shan Liu

Assignees

HEALRO CO., LTD.

Dates

Publication Date: 20260507
Application Date: 20241106

Claims (10)

1 . A data transformation system, comprising a first server, a second server and a third server that are in data communication with each other; the first server being in data communication with the second server, the first server having a data source module and a transformation processing unit, the third server being in data communication with the second server, the third server having a built-in language model, the third server being configured for collecting format descriptions of source data and data examples, in cooperation with documentation and a specification file of a standard format and a specification file of a mapping profile, through the language model, the mapping profile being created and stored in the second server; the first server extracting the source data through the data source module and transmitting it to the transformation processing unit, the transformation processing unit parsing the mapping profile and transforming the source data into the standard format according to mapping instructions after parsing.
2 . The data transformation system as claimed in claim 1 , wherein the source data is structured or unstructured data from a database, a file or by calling an application programming interface (API) in a format that is selected from the group consisting of extensible markup language (XML), JavaScript Object Notation (JSON), Resource Description Framework (RDF) and Terse RDF Triple Language (Turtle), and the transformed standard format is structured data.
3 . The data transformation system as claimed in claim 1 , wherein the second server has a user management module, the user management module provides an operation interface through which a user creates the mapping profile manually or modifies the mapping profile created by the language model, when data mapping is set through the operation interface, the transformation processing unit extracts the source data in real time and transforms the source data into the standard format according to the modified mapping profile and presents it in the operation interface in a visualized manner.
4 . The data transformation system as claimed in claim 1 , wherein the language model is first trained to understand description files in various formats, and then the language model is fine-tuned by collecting the format descriptions of the source data and the data examples, the specification file of the standard format, the specification file of the mapping profile and examples of the mapping profile file.
5 . The data transformation system as claimed in claim 4 , wherein the language model consists of a large language model (LLM), and the format descriptions of the source data are data tables or field names.
6 . The data transformation system as claimed in claim 1 , wherein transformed standard format data is exported through downloading, saving to a specified storage space or by calling an application programming interface of other systems.
7 . The data transformation system as claimed in claim 1 , further comprising a fourth server, the fourth server being in data communication with the first server, the fourth server has a database for storing transformed standard format data and a communication module for providing the standard format data through a standard protocol, so as to provide exchange and use of the standard format data.
8 . The data transformation system as claimed in claim 7 , wherein the source data is medical record data of a medical institution, the standard format data is FHIR format data of Fast Healthcare Interoperability Resources (FHIR), and the communication module is based on a FHIR protocol (FHIR RESTFul API) that implements standard data exchange to form a system framework based on the SMART App Launch Framework defined by a Health Level Seven (HL7), thereby enabling the fourth server to constitute a medical data exchange platform as a medical application.
9 . The data transformation system as claimed in claim 7 , wherein the fourth server further has a permission authorization management module that is in data communication with the communication module to perform permission management on exchange and use of the standard format data.
10 . The data transformation system as claimed in claim 1 , wherein the source data has a plurality of field items, and the transformation processing unit transforms the plurality of field items into the standard format data in real time or in batches according to the mapping profile.

Description

FIELD OF THE INVENTION The present invention relates to a data transformation system, and more particularly, to a data transformation system that can transform various formats of source data into a standard format. BACKGROUND OF THE INVENTION Data incompatibility between organizations and institutions poses a significant challenge to data integration. Variations in data formats prevent direct import into computer systems, necessitating costly and time-consuming manual programming (hard-coding) for data transformation. This approach not only increases development expenses and delays project timelines but also creates brittle, hard-to-maintain solutions that struggle to scale as data volumes and format variations grow. Taiwan Utility Model Publication No. M650536 discloses a data exchange platform using artificial intelligence algorithms for data format transformation. However, this approach still requires substantial manual effort to construct initial mapping information. Specifically, providers of source data must manually define field correspondences between their data and the target standard format. Even with advancements in generative AI, automatically inferring these complex mappings across diverse schemas remains a significant challenge due to the resource-intensive nature of these models, often requiring significant computational power and time. This high resource consumption, coupled with the need for manual intervention, prevents fully automated data transformation. Therefore, existing solutions fall short of achieving seamless conversion of various source data formats into a unified standard. SUMMARY OF THE INVENTION The primary object of the present invention is to provide a data transformation system that can transform various formats of source data into a standard format without the need to define mapping information individually. In order to achieve the foregoing object, the data transformation system provided by the present invention comprises a first server, a second server and a third server that are in data communication with each other. The first server is in data communication with the second server. The first server has a data source module and a transformation processing unit. The third server is in data communication with the second server. The third server has a built-in language model. The third server is configured for collecting format descriptions of source data and data examples, in cooperation with documentation and a specification file of a standard format and a specification file of a mapping profile. Through the language model, the mapping profile is created and stored in the second server. The first server extracts the source data through the data source module and transmits it to the transformation processing unit. The transformation processing unit parses the mapping profile and transforms the source data into the standard format according to mapping instructions after parsing. Preferably, the second server has a user management module. The user management module provides an operation interface through which a user can create the mapping profile manually or modify the mapping profile created by the language model. When data mapping is set through the operation interface, the transformation processing unit extracts the source data in real time and transforms the format of the source data into the standard format according to the modified mapping profile and presents it in the operation interface in a visualized manner. Preferably, the data transformation system further comprises a fourth server. The fourth server is in data communication with the first server. The fourth server has a database for storing transformed standard format data and a communication module for providing the standard format data through a standard protocol, so as to provide exchange and use of the standard format data. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic view of the data transformation system of the present invention; and FIG. 2 is a schematic view of the present invention when in use. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS As shown in FIG. 1 and FIG. 2, the present invention discloses a data transformation system, comprising a first server 11, a second server 21 and a third server 31 that are in data communication with each other. The term “server” referred to by the first server 11, the second server 21 and the third server 31 is a device having at least a storage unit and a communication unit. The storage unit includes, but are not limited to, random access memory (RAM), read only memory (ROM), electronically-erasable programmable read-only memory (EEPROM), flash memory, or other memory technology, compact disc read-only memory (CD-ROM), digital video disk (DVD), other optical storage, magnetic cassettes, magnetic tape, and disk storage for use in storing data. The communication unit can implement computer-readable instructions, data, structures, application modules a