Insights from machine-learned diet success prediction

Ingmar Weber, Palakorn Achananuparp

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider “quantified self“ movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model’s prediction. Our findings include both expected results, such as the token “mcdonalds” or the category “dessert” being indicative for being over the calories goal, but also less obvious ones such as the di erence between pork and poultry concerning dieting success, or the use of the “quick added calories” functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.

Original languageEnglish
Title of host publicationPacific Symposium on Biocomputing 2016, PSB 2016
PublisherWorld Scientific Publishing Co. Pte Ltd
Pages540-551
Number of pages12
Publication statusPublished - 2016
Event21st Pacific Symposium on Biocomputing, PSB 2016 - Big Island, United States
Duration: 4 Jan 20168 Jan 2016

Other

Other21st Pacific Symposium on Biocomputing, PSB 2016
CountryUnited States
CityBig Island
Period4/1/168/1/16

Fingerprint

Nutrition
Application programs
Poultry
Carbohydrates
Data mining
Learning systems
Proteins

Keywords

  • Calorie counting
  • MyFitnessPal
  • Quantified self
  • Weight loss

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Biomedical Engineering

Cite this

Weber, I., & Achananuparp, P. (2016). Insights from machine-learned diet success prediction. In Pacific Symposium on Biocomputing 2016, PSB 2016 (pp. 540-551). World Scientific Publishing Co. Pte Ltd.

Insights from machine-learned diet success prediction. / Weber, Ingmar; Achananuparp, Palakorn.

Pacific Symposium on Biocomputing 2016, PSB 2016. World Scientific Publishing Co. Pte Ltd, 2016. p. 540-551.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Weber, I & Achananuparp, P 2016, Insights from machine-learned diet success prediction. in Pacific Symposium on Biocomputing 2016, PSB 2016. World Scientific Publishing Co. Pte Ltd, pp. 540-551, 21st Pacific Symposium on Biocomputing, PSB 2016, Big Island, United States, 4/1/16.
Weber I, Achananuparp P. Insights from machine-learned diet success prediction. In Pacific Symposium on Biocomputing 2016, PSB 2016. World Scientific Publishing Co. Pte Ltd. 2016. p. 540-551
Weber, Ingmar ; Achananuparp, Palakorn. / Insights from machine-learned diet success prediction. Pacific Symposium on Biocomputing 2016, PSB 2016. World Scientific Publishing Co. Pte Ltd, 2016. pp. 540-551
@inproceedings{0862a030028949b8969ad47c7b5b90b5,
title = "Insights from machine-learned diet success prediction",
abstract = "To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider “quantified self“ movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model’s prediction. Our findings include both expected results, such as the token “mcdonalds” or the category “dessert” being indicative for being over the calories goal, but also less obvious ones such as the di erence between pork and poultry concerning dieting success, or the use of the “quick added calories” functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.",
keywords = "Calorie counting, MyFitnessPal, Quantified self, Weight loss",
author = "Ingmar Weber and Palakorn Achananuparp",
year = "2016",
language = "English",
pages = "540--551",
booktitle = "Pacific Symposium on Biocomputing 2016, PSB 2016",
publisher = "World Scientific Publishing Co. Pte Ltd",

}

TY - GEN

T1 - Insights from machine-learned diet success prediction

AU - Weber, Ingmar

AU - Achananuparp, Palakorn

PY - 2016

Y1 - 2016

N2 - To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider “quantified self“ movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model’s prediction. Our findings include both expected results, such as the token “mcdonalds” or the category “dessert” being indicative for being over the calories goal, but also less obvious ones such as the di erence between pork and poultry concerning dieting success, or the use of the “quick added calories” functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.

AB - To support people trying to lose weight and stay healthy, more and more fitness apps have sprung up including the ability to track both calories intake and expenditure. Users of such apps are part of a wider “quantified self“ movement and many opt-in to publicly share their logged data. In this paper, we use public food diaries of more than 4,000 long-term active MyFitnessPal users to study the characteristics of a (un-)successful diet. Concretely, we train a machine learning model to predict repeatedly being over or under self-set daily calories goals and then look at which features contribute to the model’s prediction. Our findings include both expected results, such as the token “mcdonalds” or the category “dessert” being indicative for being over the calories goal, but also less obvious ones such as the di erence between pork and poultry concerning dieting success, or the use of the “quick added calories” functionality being indicative of over-shooting calorie-wise. This study also hints at the feasibility of using such data for more in-depth data mining, e.g., looking at the interaction between consumed foods such as mixing protein- and carbohydrate-rich foods. To the best of our knowledge, this is the first systematic study of public food diaries.

KW - Calorie counting

KW - MyFitnessPal

KW - Quantified self

KW - Weight loss

UR - http://www.scopus.com/inward/record.url?scp=84966623765&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84966623765&partnerID=8YFLogxK

M3 - Conference contribution

C2 - 26776216

AN - SCOPUS:84966623765

SP - 540

EP - 551

BT - Pacific Symposium on Biocomputing 2016, PSB 2016

PB - World Scientific Publishing Co. Pte Ltd

ER -