Predicting missing items

Patent metadata

Field	Value
Publication number	US-11488107-B2
Application number	US-201916708000-A
Country	US
Kind code	B2
Filing date	Dec 9, 2019
Priority date	Dec 9, 2019
Publication date	Nov 1, 2022
Grant date	Nov 1, 2022

Abstract

In some embodiments, there is provided a system. The system may include at least one data processor and at least one memory storing instructions which, when executed by the at least one data processor, cause the apparatus to at least: determine, for a received document including at least one item, that the received document likely includes at least one missing item, the determination based on at least a machine learning model and the at least one item; and provide an indication of the at least one missing item. Related systems and articles of manufacture are also provided.

First claim

What is claimed is:

1. A system comprising:
at least one data processor; and
at least one memory storing instructions which, when executed by the at least one data processor, causes the system to at least:
train, using at least a set of reference documents, a machine-learning model comprising a co-occurrence model, the set of reference documents each including a set of items verified to confirm the corresponding set of items is complete and is not missing any items, the set of items corresponding to hospital billing codes, and wherein the trained machine learning model comprising the co-occurrence model is trained to detect one or more missing hospital billing codes;
receive a document comprising at least a first hospital billing code;
determine, for the received document including the first hospital billing code, that the received document is missing at least a second hospital billing code, the determination based on at least the trained machine learning model and at least the first hospital billing code provided as an input to the trained machine learning model, the trained machine-learning model comprising the co-occurrence model providing a likelihood that the second hospital billing code is missing from the received document that includes the first hospital billing code, the co-occurrence model comprising a matrix including values representative of likelihoods that pairs of hospital billing codes are likely to be included in the received document, the pairs including the first hospital billing code and the missing second hospital billing code; and

provide an indication of at least the missing second hospital billing code, wherein the indication comprises a recommendation to add the missing second hospital billing code to the received document.

2. The system of claim 1 , wherein the system is further caused to at least:
provide the likelihood as a confidence value that the second hospital billing code is missing from the received document.

3. The system of claim 2 , wherein the system is further caused to at least:
provide the likelihood as a score.

4. A method comprising:
training, using at least a set of reference documents, a machine-learning model comprising a co-occurrence model, the set of reference documents each including a set of items verified to confirm the corresponding set of items is complete and is not missing any items, the set of items corresponding to hospital billing codes, and wherein the trained machine learning model comprising the co-occurrence model is trained to detect one or more missing hospital billing codes;
receiving a document comprising at least a first hospital billing code;
determining, for the received document including the first hospital billing code, that the received document is missing at least a second hospital billing code, the determination based on at least the trained machine learning model and at least the first hospital billing code provided as an input to the trained machine learning model, the trained machine-learning model comprising the co-occurrence model providing a likelihood that the second hospital billing code is missing from the received document that includes the first hospital billing code, the co-occurrence model comprising a matrix including values representative of likelihoods that pairs of hospital billing codes are likely to be included in the received document, the pairs including the first hospital billing code and the missing second hospital billing code; and
providing an indication of at least the missing second hospital billing code, wherein the indication comprises a recommendation to add the missing second hospital billing code to the received document.

5. The method of claim 4 , further comprising providing the likelihood as a confidence value that the second hospital billing code is missing from the received document.

6. The method of claim 4 , wherein the likelihood is provided as a score.

7. A non-transitory computer-readable storage medium including program code which when executed causes operations comprising:
training, using at least a set of reference documents, a machine-learning model comprising a co-occurrence model, the set of reference documents each including a set of items verified to confirm the corresponding set of items is complete and is not missing any items, the set of items corresponding to hospital billing codes, and wherein the trained machine learning model comprising the co-occurrence model is trained to detect one or more missing hospital billing codes;
receiving a document comprising at least a first hospital billing code;
determining, for the received document including the first hospital billing code, that the received document is missing at least a second hospital billing code, the determination based on at least the trained machine learning model and at least the first hospital billing code provided as an input to the trained machine learning model, the trained machine-learning model comprising the co-occurrence model providing a likelihood that the second hospital billing code is missing from the received document that includes the first hospital billing code, the co-occurrence model comprising a matrix including values representative of likelihoods that pairs of hospital billing codes are likely to be included in the received document, the pairs including the first hospital billing code and the missing second hospital billing code; and
providing an indication of at least the missing second hospital billing code, wherein the indication comprises a recommendation to add the missing second hospital billing code to the received document.

Assignees

• Sap Se

Inventors

• Toldo Luca
• Schenk Konrad
• Atilgan Tunahan
• Hoecker Martin
• Lieske Bettina
• Jiang Ying

Classifications

• G06F18/2321

  using statistics or function optimisation, e.g. modelling of probability density functions · CPC title

• G06F18/24137

  Distances to cluster centroïds · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06V30/413

  Classification of content, e.g. text, photographs or tables · CPC title

• G06Q10/0875 Primary

  Itemisation or classification of parts, supplies or services, e.g. bill of materials · CPC title