scholarly journals Comparison of cluster and principal component analysis techniques to derive dietary patterns in Irish adults

2008 ◽  
Vol 101 (4) ◽  
pp. 598-608 ◽  
Author(s):  
Áine P. Hearty ◽  
Michael J. Gibney
Keyword(s):  
Principal Component Analysis ◽  
Cluster Analysis ◽  
Energy Intake ◽  
Dietary Patterns ◽  
Principal Component ◽  
Food Group ◽  
Food Groups ◽  
Healthy Foods ◽  
Percentage Contribution ◽  
Group Variable

The aims of the present study were to examine and compare dietary patterns in adults using cluster and factor analyses and to examine the format of the dietary variables on the pattern solutions (i.e. expressed as grams/day (g/d) of each food group or as the percentage contribution to total energy intake). Food intake data were derived from the North/South Ireland Food Consumption Survey 1997–9, which was a randomised cross-sectional study of 7 d recorded food and nutrient intakes of a representative sample of 1379 Irish adults aged 18–64 years. Cluster analysis was performed using thek-means algorithm and principal component analysis (PCA) was used to extract dietary factors. Food data were reduced to thirty-three food groups. For cluster analysis, the most suitable format of the food-group variable was found to be the percentage contribution to energy intake, which produced six clusters: ‘Traditional Irish’; ‘Continental’; ‘Unhealthy foods’; ‘Light-meal foods & low-fat milk’; ‘Healthy foods’; ‘Wholemeal bread & desserts’. For PCA, food groups in the format of g/d were found to be the most suitable format, and this revealed four dietary patterns: ‘Unhealthy foods & high alcohol’; ‘Traditional Irish’; ‘Healthy foods’; ‘Sweet convenience foods & low alcohol’. In summary, cluster and PCA identified similar dietary patterns when presented with the same dataset. However, the two dietary pattern methods required a different format of the food-group variable, and the most appropriate format of the input variable should be considered in future studies.

scholarly journals Dietary patterns in Irish adolescents: a comparison of cluster and principal component analyses

2011 ◽  
Vol 16 (5) ◽  
pp. 848-857 ◽  
Author(s):  
Áine P Hearty ◽  
Michael J Gibney
Keyword(s):  
Cluster Analysis ◽  
Total Energy ◽  
Dietary Patterns ◽  
Principal Component ◽  
Food Groups ◽  
Food Diary ◽  
Breakfast Cereal ◽  
Percentage Contribution ◽  
Main Meal ◽  
The Impact

AbstractObjectivePattern analysis of adolescent diets may provide an important basis for nutritional health promotion. The aims of the present study were to examine and compare dietary patterns in adolescents using cluster analysis and principal component analysis (PCA) and to examine the impact of the format of the dietary variables on the solutions.DesignAnalysis was based on the Irish National Teens Food Survey, in which food intake data were collected using a semi-quantitative 7 d food diary. Thirty-two food groups were created and were expressed as either g/d or percentage contribution to total energy. Dietary patterns were identified using cluster analysis (k-means) and PCA.SettingRepublic of Ireland, 2005–2006.SubjectsA representative sample of 441 adolescents aged 13–17 years.ResultsFive clusters based on percentage contribution to total energy were identified, ‘Healthy’, ‘Unhealthy’, ‘Rice/Pasta dishes’, ‘Sandwich’ and ‘Breakfast cereal & Main meal-type foods’. Four principal components based on g/d were identified which explained 28 % of total variance: ‘Healthy foods’, ‘Traditional foods’, ‘Sandwich foods’ and ‘Unhealthy foods’.ConclusionsA ‘Sandwich’ and an ‘Unhealthy’ pattern are the main dietary patterns in this sample. Patterns derived from either cluster analysis or PCA were comparable, although it appears that cluster analysis also identifies dietary patterns not identified through PCA, such as a ‘Breakfast cereal & Main meal-type foods’ pattern. Consideration of the format of the dietary variable is important as it can directly impact on the patterns obtained for both cluster analysis and PCA.

scholarly journals The reproducibility and stability of dietary patterns

2020 ◽  
Vol 79 (OCE2) ◽  
Author(s):  
Orla Prendiville ◽  
Aoife E. McNamara ◽  
Lorraine Brennan
Keyword(s):  
Cluster Analysis ◽  
Dietary Patterns ◽  
Dietary Pattern ◽  
Food Group ◽  
Multiple Time ◽  
Food Groups ◽  
Energy Contribution ◽  
Dietary Analysis ◽  
Breakfast Cereals ◽  
Time Points

AbstractA person's dietary intake consists of multiple foods eaten as part of a meal as opposed to any one single food/nutrient. Therefore, it is important to understand the interactions between foods and how they affect diet-disease associations. As a result, dietary patterns have emerged as important tools in nutrition research. The objective of the current study is to assess the reproducibility and stability of dietary patterns across four different time-points. Anthropometric measurements were taken from a subset of participants of a free-living cohort study (n = 94), followed by the administration of a 24-hour dietary recall once a month, for four months. The dietary data was entered into dietary analysis software, Nutritics, by two researchers independently, and cross-checked. Foods were assigned to one of 33 predefined food groups, which were further collapsed to 18 food groups based on previous research. Statistical analysis was then performed on the final dataset. Intra-class correlation coefficients were derived to assess the reproducibility of each food group across the four time-points. Variables were standardized using z-scores and dietary patterns were derived using K-means cluster analysis. Stability was assessed by coding participants into one of six groups based on their dietary pattern transition between visit one and four. Analysis of this sub cohort revealed that the intake of food groups (% energy contribution) was reproducible across the time-points. The majority had good to very good agreement, with vegetables and vegetable dishes having the strongest agreement (ICC = 0.831) followed by milk and yogurts (ICC = 0.773), fruit and fruit dishes (ICC = 0.729), and breakfast cereals (ICC = 0.680). Two distinct dietary patterns were identified at each time-point; a ‘Healthy’ and an ‘Unhealthy’ dietary pattern. The ‘Healthy’ dietary pattern was characterized by a significantly higher energy contribution (p < 0.05) from the following food groups – vegetables and vegetable dishes; fruit and fruit dishes; milk and yogurts; breakfast cereals; butter, spreading fats and oils. The analysis on stability demonstrated 42% of participants remained in the same dietary pattern, while 58% transitioned from one dietary pattern to the other. Our results to date demonstrate that two distinct dietary patterns can be derived across multiple time-points using cluster analysis and the food group composition of these dietary patterns can be considered reproducible. Future work will explore these dietary patterns further incorporating the entire cohort and linking stability to health parameters.

Abstract P075: Five Dietary Patterns Emerge in the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study

Circulation ◽  
2012 ◽  
Vol 125 (suppl_10) ◽  
Author(s):  
Abraham J Letter ◽  
Suzanne E Judd ◽  
James M Shikany ◽  
David L Roth ◽  
P. K Newby
Keyword(s):  
Factor Analysis ◽  
Racial Differences ◽  
Dietary Patterns ◽  
Fruits And Vegetables ◽  
Principal Component ◽  
Food Group ◽  
Food Groups ◽  
Black And White ◽  
Southeastern Us ◽  
Regards Study

Introduction: Examining diet as a whole using dietary patterns methods rather than focusing on single food group or nutrient exposures may be more informative when studying associations of diet and disease. Several large cohorts have used factor analysis to empirically derive dietary patterns but few have employed a cohort as geographically and racially diverse as the REGARDS study. Methods: The REGARDS study is a cohort of 30,239 Black and White adults age 45 and older, half of whom reside in the Southeastern US (also known as the stroke belt); the remainder reside elsewhere in the continental US. The present analysis included 21,636 participants who completed the Block 98 food frequency questionnaire at baseline. Principal Component Analysis (PCA) with varimax rotation was used on a split sample to determine dietary patterns based on 56 food groups. Race, sex, and region-specific solutions were evaluated for congruence alongside scree plots, eigenvalues, and interpretability. Confirmatory Factor Analysis (CFA) was utilized on the second half of the sample for validation of the PCA findings. Results: Sub-group analyses showed acceptable congruence and interpretability, thus we performed PCA on the entire sample in the final solution. Five dietary patterns emerged: the “traditional” pattern was characterized by mixed dishes; the “healthy” pattern by fruits and vegetables; the “sweets” pattern by sweet snacks and desserts; the “Southern” pattern by fried food, organ meat, and sweetened beverages; and the “alcohol” pattern by beer, wine, liquor, and salads. There were marked differences in factor score means across demographic and socioeconomic groups. For example, Blacks were much more likely than Whites to have a Southern diet. Discussion: Clear and meaningful dietary patterns emerged in this large cohort of Black and White Americans. Variability in dietary intake across demographic factors emphasizes the need to explore how these factors contribute to differential susceptibility to stroke and other chronic diseases.

scholarly journals Education and lifestyle predict change in dietary patterns and diet quality of adults 55 years and over

2019 ◽  
Vol 18 (1) ◽  
Author(s):  
Maree G. Thorpe ◽  
Catherine M. Milte ◽  
David Crawford ◽  
Sarah A. McNaughton
Keyword(s):  
Physical Activity ◽  
Older Adults ◽  
Principal Component Analysis ◽  
Diet Quality ◽  
Dietary Patterns ◽  
Principal Component ◽  
Component Analysis ◽  
Processed Meat ◽  
Food Groups ◽  
Time Points

Abstract Background Diet is a key risk factor for chronic disease, and an increasing concern among older adults. We aim to examine the changes in dietary patterns using principal component analysis and a diet quality index among older adults and examine the predictors of dietary change over a 4 year period. Methods Data was obtained via a postal survey in a prospective cohort, the Wellbeing Eating and Exercise for a Long Life (WELL) study. Australian adults aged 55 years and over (n = 1005 men and n = 1106 women) completed a food frequency at three time points and provided self-reported personal characteristics. Principal component analysis was used to assess dietary patterns and diet quality was assessed using the 2013 Revised Dietary Guideline Index. The relationships between predictors and change in dietary patterns were assessed by multiple linear regression. Results Two dietary patterns were consistently identified in men and women at three time points over 4 years. One was characterised by vegetables, fruit and white meat, and the other was characterised by red and processed meat and processed foods. Reduced consumption of key food groups within the principal component analysis-determined dietary patterns was observed. An increase in diet quality over 4 years was observed in men only. Reported higher education levels and favourable lifestyle characteristics, including not smoking and physical activity, at baseline predicted an increase in healthier dietary patterns over 4 years. Conclusions There was stability in the main dietary patterns identified over time, however participants reported an overall decrease in the frequency of consumption of key food groups. Compliance with the Australian Dietary Guidelines remained poor and therefore targeting this population in nutritional initiatives is important. Design of nutrition promotion for older adults need to consider those with lower socioeconomic status, as having a lower level of education was a predictor of poorer dietary patterns. It is important to consider how nutrition behaviours can be targeted alongside other lifestyle behaviours, such as smoking and inadequate physical activity to improve health.

scholarly journals Meal patterns associated with energy intake in people with obesity

2021 ◽  
pp. 1-48
Author(s):  
Cathrine Horn ◽  
Johnny Laupsa-Borge ◽  
Amanda I. O. Andersen ◽  
Laurence Dyer ◽  
Ingrid Revheim ◽  
...  
Keyword(s):  
Cluster Analysis ◽  
Energy Intake ◽  
Dietary Patterns ◽  
Fast Food ◽  
Time Of Day ◽  
Activity Level ◽  
Eating Patterns ◽  
Food Groups ◽  
Anthropometric Measures ◽  
Meal Patterns

Abstract It is widely assumed that people with obesity have several common eating patterns, including breakfast-skipping (1), eating during the night (2) and high fast-food consumption (3). However, differences in individual meal and dietary patterns may be crucial to optimizing obesity treatment. Therefore, we investigated the inter-individual variation in eating patterns, hypothesizing that individuals with obesity show different dietary and meal patterns, and that these associate with self-reported energy intake (rEI) and/or anthropometric measures. Cross-sectional data from 192 participants (aged 20–55 years) with obesity, including 6 days of weighed food records, were analyzed. Meal patterns and dietary patterns were derived using exploratory hierarchical cluster analysis and k-means cluster analysis, respectively. Five clear meal patterns were found based on the time-of-day with the highest mean rEI. The daily rEI (mean ± SD kcal) was highest among “midnight-eaters” (2550 ± 550), and significantly (p < 0.05) higher than “dinner-eaters” (2060 ± 550), “lunch-eaters” (2080 ± 520), and “supper-eaters” (2100 ± 460), but not “regular-eaters” (2330 ± 650). Despite differences of up to 490 kcal between meal patterns, there were no significant differences in anthropometric measures or physical activity level (PAL). Four dietary patterns were also found with significant differences in intake of specific food groups, but without significant differences in anthropometry, PAL, or rEI. Our data highlight meal timing as a determinant of individual energy intake in people with obesity. The study supports the importance of considering a person’s specific meal pattern, with possible implications for more person-focused guidelines and targeted advice.

scholarly journals Dietary patterns in a nursing team measured by principal component analysis

2020 ◽  
Vol 54 ◽  
Author(s):  
Patricia Lima Dias Barreiro ◽  
Ana Glória Godoi Vasconcelos ◽  
Lucia Rotenberg ◽  
Rosane Harter Griep ◽  
Odaleia Barbosa de Aguiar
Keyword(s):  
Principal Component Analysis ◽  
Dietary Patterns ◽  
Principal Component ◽  
Component Analysis ◽  
Nursing Assistants ◽  
Food Groups ◽  
Food Pattern ◽  
Analysis Technique ◽  
Alcoholic Drinks ◽  
Nursing Professionals

Abstract Objective: To characterize the dietary pattern of nursing professionals at a public hospital in Rio de Janeiro, RJ, Brazil. Method: A sectional study with nursing professionals (nurses, technicians and nursing assistants). Two 24-hour food recall records were applied, totaling 459 foods, being reduced to 24 food groups. Food patterns were identified using the Principal Component Analysis technique, followed by orthogonal varimax rotation. A Scree Plot graph indicated three factors to be extracted and loads > +0.30 were adopted to define dietary patterns. Results: A total of 309 professionals participated. The sample consisted of 85.8% of female individuals. The patterns were named “traditional” which included rice (0.747), beans (0.702) and meat (0.713); “healthy”: vegetables (0.444), greens (0.450), fruits (0.459), bananas and oranges (0.379), and “snacks”: sugar (0.661), bread (0.471), cakes and cookies (0.334), non-alcoholic drinks (0.727). Conclusion: The results highlight the “traditional” food pattern of Brazilian food consumption based on the combination of rice, beans and meat. Future studies may investigate the effect of dietary patterns on health outcomes among nursing workers.

scholarly journals Dietary pattern changes in Fukushima residents after the Great East Japan Earthquake: the Fukushima Health Management Survey 2011–2013

2020 ◽  
pp. 1-10
Author(s):  
Enbo Ma ◽  
Tetsuya Ohira ◽  
Hironori Nakano ◽  
Masaharu Maeda ◽  
Hirooki Yabe ◽  
...  
Keyword(s):  
Dietary Patterns ◽  
Health Management ◽  
Short Form ◽  
Correlation Coefficients ◽  
Principal Component ◽  
Food Group ◽  
Great East Japan Earthquake ◽  
Food Groups ◽  
Eating Behaviours ◽  
Typical Pattern

Abstract Objective: Dietary patterns more closely resemble actual eating behaviours because multiple food groups, not a single food group or nutrient, are considered. The present study aimed to identify and assess changes of dietary patterns in Fukushima residents. Design: Dietary data were collected using a short-form FFQ in annual Fukushima Health Management Survey between 2011 and 2013 after the Great East Japan Earthquake. Year- and sex-specific dietary patterns were determined by the principal component analysis. Setting: Evacuation and nonevacuation zones in Fukushima, Japan. Participants: Eligible participants aged ≥16 years answered the FFQ (n 67 358 in 2011, n 48 377 in 2012 and n 40 742 in 2013). Results: Three identified dietary patterns were assessed similarly in men and women and among years: typical, juice and meat. In total participants, the Spearman’s correlation coefficients between two survey years were 0·70–0·74 for the typical, 0·58–0·66 for the juice and 0·50–0·54 for the meat pattern scores. Adjusted for sociodemographic factors, evacuees had lower typical pattern scores, higher juice pattern scores and the same meat pattern scores compared with non-evacuees. The means of typical pattern scores in evacuees and it of juice pattern scores in non-evacuees continued declining over years. Similar profiles of dietary patterns and trends of pattern scores were observed in participants (n 22 805) who had provided three dietary assessments. Conclusions: Changes of dietary patterns have been observed between 2011 and 2013. Careful investigation of those with low intake of typical pattern foods and promotion of them, particularly in evacuees, are needed.

Classification of Observations through Combination of the Dimension Reduction and the Cluster Analysis

2017 ◽  
Vol 7 (8) ◽  
pp. 30
Author(s):  
Hyeuk Kim
Keyword(s):  
Machine Learning ◽  
Principal Component Analysis ◽  
Cluster Analysis ◽  
Unsupervised Learning ◽  
Principal Component ◽  
Component Analysis ◽  
Baseball Players ◽  
Partitioning Around Medoids ◽  
Different Characteristics

Unsupervised learning in machine learning divides data into several groups. The observations in the same group have similar characteristics and the observations in the different groups have the different characteristics. In the paper, we classify data by partitioning around medoids which have some advantages over the k-means clustering. We apply it to baseball players in Korea Baseball League. We also apply the principal component analysis to data and draw the graph using two components for axis. We interpret the meaning of the clustering graphically through the procedure. The combination of the partitioning around medoids and the principal component analysis can be used to any other data and the approach makes us to figure out the characteristics easily.

Sign in / Sign up

Export Citation Format

Share Document