Semantics for Personal Health


This website contains illustrative applications of semantic technologies for enabling clinically relevant personal health applications using nutrition behavior as a focus.

Find Out More

Semantics for Personal Health


We develop a knowledge model called the Personal Health Ontology, which is grounded in semantic technologies, and represents a patient's combined medical information, social determinants of health, and observations of daily living elicited from interviews with diabetic patients. We then generate a personal health knowledge graph that captures temporal patterns from synthesized food logs, annotated with the concepts from the Personal Health Ontology. We further discuss how lifestyle guidelines grounded in semantic technologies can be reasoned with the generated personal health knowledge graph to provide appropriate dietary recommendations that satisfy the user's medical and other lifestyle needs.

Get Started!

Personal Health Ontology


Driven by Real-World Data

Based on a set of interviews conducted with 21 people who declared themselves to be within five years of being diagnosed with T2D, we developed the ontology to characterize the dietary behaviors of diabetes patients.

Personal Health Knowledge Graph


Synthetic Meal-Level Food Log Dataset

We used synthetic food log data at the meal-level granularity (i.e., breakfast, lunch, snack, and dinner), which spans five weeks. This dataset follows the MyFitnessPal schema, offering the nutrient information of the foods consumed by the user, as well as the food names.

Guideline-Influenced Insight Discovery

We extended a version of an existing time-series summarization (TSS) framework to mine synthetic food log data for relevant behavioral insights. These insights were inspired by a selected set of dietary guidelines defined by the American Diabetes Association for diabetics and pre-diabetics.

RDF Triple Generation for PHKG Population

Using concepts and relationships defined within the Personal Health Ontology, we translate the discovered temporal insights into RDF triples. These triples are, then, used to populate the Personal Health Knowledge Graph along with personal information about the user and structured food history data.

Timely Repopulation for Constraint Queries

As we are working with temporal personal and health information, the Personal Health Knowledge Graph is expected to update its constraints in a timely manner to provide the most accurate and relevant constraints when it is queried. To achieve this, it is repopulated by the TSS framework.

Recommendation Engine


Relevant Papers


Let's Get In Touch!


Please use the form below to contact us if you have any questions.