Tapio Nummi | Tampere University (original) (raw)
Papers by Tapio Nummi
Journal of Multivariate Analysis, Jul 1, 2022
In the process of modeling longitudinal data, we focus on the case that the studied population is... more In the process of modeling longitudinal data, we focus on the case that the studied population is comprised of different groups of individuals and individuals within the same group share the similar kind of mean progression trajectories, where finite mixture models (FMM) are often used to address this kind of unobserved heterogeneity in terms of mean. Existing methods, such as parametric and semiparametric mixture regression, usually model the mean in each subpopulation with assumption that observations sharing a common trajectory are independent or their covariance structure is pre-specified, but less research considers modeling of covariance structures while accounting for heterogeneity. In this paper, we introduce a joint model which models the mean and covariance structures simultaneously in a finite normal mixture regression, demonstrating how important the within-subject correlation is in clustering longitudinal data. Model parameters are estimated with an iteratively re-weighted least squares EM algorithm (IRLS-EM). Our estimators are shown to be consistent and asymptotically normal. We can identify different mean trajectories and covariance structures in all clusters. Simulations show that the proposed method performs well and gives more accurate clustering results by introducing covariance modeling. Real data analysis is also used to illustrate the usefulness of the proposed method, and it presents good performance in clustering COVID-19 deaths for European countries in terms of progression trajectory.
International Journal of Environmental Research and Public Health, Oct 5, 2022
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
Journal of Applied Statistics, 2004
The aim of this study was to investigate prediction of stem measurements of Scots pine(Pinus sylv... more The aim of this study was to investigate prediction of stem measurements of Scots pine(Pinus sylvestris L.) for a modern computerized forest harvester. We are interested in the prediction of stem curve measurements when measurements of stems already processed and a short section of the stem under process are known. The techniques presented here are based on cubic smoothing splines and on multivariate regression models. One advantage of these methods is that they do not assume any special functional form of the stem curve. They can also be applied to the prediction of branch limits and stem height of pine stems.
Acta Paediatrica, Jun 8, 2016
We tracked the body mass index (BMI) of children born to mothers with or without gestational diab... more We tracked the body mass index (BMI) of children born to mothers with or without gestational diabetes mellitus (GDM) or type 1 diabetes from birth to 12 years of age and examined the trends in both diseases. Methods: Antenatal and postnatal health survey data were collected from 6909 Finnish children born at six time points between 1974 and 2004. We compared the BMI trajectory between the offspring of mothers with and without GDM or type 1 diabetes, and the association between GDM and overweight offspring was analysed. Results: The prevalence of GDM and type 1 diabetes increased markedly over the study period. The BMI trajectory in the GDM or type 1 diabetes offspring differed significantly from the nondiabetic offspring. The timing of adiposity rebound occurred significantly earlier in the GDM (4.8 years) and type 1 diabetes (4.4 years) groups than the nondiabetic group (5.5 years). GDM offspring were more likely to be overweight at five, seven and 12 years of age (24.6%, 28.1%, 29.4%) than nondiabetic offspring (15.6%, 18.3%, 18.1%). Conclusion: Children born to mothers with GDM were significantly more likely to be overweight at an early age than those born to nondiabetic mothers. Abbreviations BMI, Body mass index; GDM, Gestational diabetes mellitus.
Työelämän tutkimus, Nov 1, 2007
arXiv (Cornell University), Sep 19, 2022
We investigate a novel non-parametric regression-based clustering algorithm for longitudinal data... more We investigate a novel non-parametric regression-based clustering algorithm for longitudinal data analysis. Combining natural cubic splines with Gaussian mixture models (GMM), the algorithm can produce smooth cluster means that describe the underlying data well. However, there are some shortcomings in the algorithm: high computational complexity in the parameter estimation procedure and a numerically unstable variance estimator. Therefore, to further increase the usability of the method, we incorporated approaches to reduce its computational complexity, we developed a new, more stable variance estimator, and we developed a new smoothing parameter estimation procedure. We show that the developed algorithm, SMIXS, performs better than GMM on a synthetic dataset in terms of clustering and regression performance. We demonstrate the impact of the computational speed-ups, which we formally prove in the new framework. Finally, we perform a case study by using SMIXS to cluster vertical atmospheric measurements to determine different weather regimes.
European journal of public health, Oct 20, 2017
an increase of 5.2 percentage points (95% CI 4.8-5.6); among males, the increase was 2.7 percenta... more an increase of 5.2 percentage points (95% CI 4.8-5.6); among males, the increase was 2.7 percentage points into 8.9% (95% CIs 2.4-2.9, 8.5%-9.4%). Adjusted hazard ratios of SSU, contrasting the cohorts, were 1.61 (95% CI 1.50-1.72) and 1.47 (95% CI 1.35-1.60) for females and males respectively. With the exception of schizophrenia, increased incidence was observed for all outcomes: hazard ratio estimates ranged from 1.15 to 6.11. Conclusions An increasing proportion of young people are coming to contact with specialized mental health related services during adolescence. Regardless of gender, this increase is evident across the varied spectrum of psychiatric and neurodevelopmental disorders and self-harm.
American Journal of Men's Health, Jul 14, 2009
http://ajmh.sagepub.com D espite the fact that fatherhood and father involvement in parenthood ha... more http://ajmh.sagepub.com D espite the fact that fatherhood and father involvement in parenthood have emerged as prominent areas of family research during the past decade (Lamb & Michael, 2000; Marsiglio, Amato, Day, & Lamb, 2000), the matter of father presence has attracted little attention in nursing and health care science. Social and behavioral scientists share the view that father's presence promotes the child's social,
Scandinavian Journal of Rheumatology, Aug 30, 2016
Objectives: To evaluate safety and effects of a new home treatment method, a whole-body cold mist... more Objectives: To evaluate safety and effects of a new home treatment method, a whole-body cold mist treatment, on chronic inflammatory arthritis patients. Method: Whole-body cold mist shower therapy was given to 121 voluntary patients with chronic inflammatory arthritis in this cross-over study during one week rehabilitation periods. Pain and sleep quality were assessed by a ten centimeter visual analogue scale (VAS). Mental status was assessed by DEPS-depression score. Body temperature, blood pressure, heart rate, use of occasional pain and sleep medication, and possible side effects were recorded. Results: The differences in pain (VAS) between treatment and control periods were significant (2.0 vs 2.4, p=0.006, paired t-test) in the last measurement, when assessed the pain of the passed week as a whole. A trend could be seen to an increasing difference towards the end of the week. The treatment effect was statistically significant (LRT, p<0.0001) after controlling for period and sequence effects. There was an indication for a better sleep quality (VAS) during the treatment period (2.3 vs 2.7, p=0.058 paired t-test) when assessing the passed week as a whole. The mean depression scale (DEPS) values showed no difference between the periods (5.5 vs 5.0, p= 0.1874 paired t-test, at start, and 4.5 vs 4.1 p=0.29 paired t-test, at the end). No significant side effects were recorded. Conclusions: The new whole-body cold treatment method may offer a safe option for pain selftreatment at home, but further study is needed to find out the clinical significance of the effect after a longer use.
European Journal of Public Health
Background Somatic symptoms among adolescents are common, yet little is known about long-term tra... more Background Somatic symptoms among adolescents are common, yet little is known about long-term trajectories of somatic symptoms and the factors in adolescence that shape them. We examined individual, family and school-based factors at age 16 as predictors of trajectories of somatic symptoms over 27 years. Methods Participants from the Northern Swedish Cohort (n = 1001) responded to questions about individual factors (e.g. health behaviours), family factors (e.g. contact with parents, social and material adversity) and school satisfaction at age 16; as well as 10 somatic symptoms at ages 16, 18, 21, 30 and 43. Teacher assessments at age 16 included overall ability at school and peer relations. Age 16 predictors of somatic symptom trajectory group membership were analysed using multinomial logistic regression. Results Poor contact with mother and poor school satisfaction were significant predictors of adverse symptom trajectories among both men and women. Low birth weight and low paren...
Data Science and SDGs, 2021
Transfer: European Review of Labour and Research, 2018
The connection between high-involvement management (HIM), entailing heavy employee involvement, a... more The connection between high-involvement management (HIM), entailing heavy employee involvement, and employee well-being is a controversial and widely discussed topic. Clarifying how job satisfaction and stress are connected to HIM and job control (the control employees have over their work), this study is based on data from two Finnish sources: an employer survey investigating the extent of HIM within an organisation, and employee assessments of job control, stress and job satisfaction. Logistic regression models were used as the study method. In contrast to previous Finnish studies, our findings show that HIM seems hardly to benefit employee well-being. Especially in the public sector, the correlation between extensive HIM and employee well-being turned out to be negative. However, HIM in the private sector was positively related to job satisfaction. As expected, a high level of job control was regularly associated with greater well-being.
Journal of Youth Studies, 2020
This study explores young Finns' transition pathways into education and work over a nine-year per... more This study explores young Finns' transition pathways into education and work over a nine-year period (from the age of 18 to 26), with a special focus on difficulties with labour market attachment. The data were derived from national registers comprising all men (29,383) and women (27,784) born in 1987. Their labour market attachment was investigated using trajectory analysis with four variables (employment, education, unemployment, social security allowances) from 2005 to 2013. We identified ten trajectory groups, of which seven in men and eight in women were categorised as 'mainstream paths', reflecting relatively smooth transitions. The other trajectory groups were considered to assume 'straying paths': 11 per cent of both men and women followed these paths. The size and nature of these paths differed between men and women: it seems that men more often have serious difficulties with labour market attachment than women do. Overall, the results of the multiple variable trajectory analyses suggest that cross-sectional inspections of the Finnish register data overestimate the proportion of NEETs, particularly among men.
One of the most popular methods for modelling survival analysis data is the ubiquitous and time-h... more One of the most popular methods for modelling survival analysis data is the ubiquitous and time-honored proportional hazards model of Cox (1972). Popularity notwithstanding, in several cases, however, the proportional hazards assumption is found to be violated. Thus, important model extensions have been developed in the intervening years. One such extension is the so-called Frailty model (see for example Collett, 2015), which is based on the utilization of random effects. Another extension arises from the application of finite mixtures in the context of time-to-event data. Although finite mixture modelling tools are used in many fields of science, they have been less well-developed and used in survival analysis. Thus, a key aim of this article is to provide an interesting application of mixture modeling in survival analysis and to discuss aspects arising in its application. In this paper, we apply these techniques using real data from the research register of the Finnish Centre for Pensions; using pension insurance mortality data, we use the basic Cox proportional hazards model by incorporating finite mixture modelling techniques. Additional comparisons with frailty models are also provided.
Social adversities in adolescence predict unfavourable trajectories of internalized mental health... more Social adversities in adolescence predict unfavourable trajectories of internalized mental health symptoms
Response to reviewers Reviewer 1:Anu Molarius Reviewer's report: TThe authors have sufficiently r... more Response to reviewers Reviewer 1:Anu Molarius Reviewer's report: TThe authors have sufficiently replied to almost all of the comments raised in the review and revised the manuscript accordingly. I have two further comments that do not necessarily require changes in the manucsript. Response 1:1. Thank you for this.
Data Science and SDGs, 2021
New Advances in Statistics and Data Science, 2017
Journal of Multivariate Analysis, Jul 1, 2022
In the process of modeling longitudinal data, we focus on the case that the studied population is... more In the process of modeling longitudinal data, we focus on the case that the studied population is comprised of different groups of individuals and individuals within the same group share the similar kind of mean progression trajectories, where finite mixture models (FMM) are often used to address this kind of unobserved heterogeneity in terms of mean. Existing methods, such as parametric and semiparametric mixture regression, usually model the mean in each subpopulation with assumption that observations sharing a common trajectory are independent or their covariance structure is pre-specified, but less research considers modeling of covariance structures while accounting for heterogeneity. In this paper, we introduce a joint model which models the mean and covariance structures simultaneously in a finite normal mixture regression, demonstrating how important the within-subject correlation is in clustering longitudinal data. Model parameters are estimated with an iteratively re-weighted least squares EM algorithm (IRLS-EM). Our estimators are shown to be consistent and asymptotically normal. We can identify different mean trajectories and covariance structures in all clusters. Simulations show that the proposed method performs well and gives more accurate clustering results by introducing covariance modeling. Real data analysis is also used to illustrate the usefulness of the proposed method, and it presents good performance in clustering COVID-19 deaths for European countries in terms of progression trajectory.
International Journal of Environmental Research and Public Health, Oct 5, 2022
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
Journal of Applied Statistics, 2004
The aim of this study was to investigate prediction of stem measurements of Scots pine(Pinus sylv... more The aim of this study was to investigate prediction of stem measurements of Scots pine(Pinus sylvestris L.) for a modern computerized forest harvester. We are interested in the prediction of stem curve measurements when measurements of stems already processed and a short section of the stem under process are known. The techniques presented here are based on cubic smoothing splines and on multivariate regression models. One advantage of these methods is that they do not assume any special functional form of the stem curve. They can also be applied to the prediction of branch limits and stem height of pine stems.
Acta Paediatrica, Jun 8, 2016
We tracked the body mass index (BMI) of children born to mothers with or without gestational diab... more We tracked the body mass index (BMI) of children born to mothers with or without gestational diabetes mellitus (GDM) or type 1 diabetes from birth to 12 years of age and examined the trends in both diseases. Methods: Antenatal and postnatal health survey data were collected from 6909 Finnish children born at six time points between 1974 and 2004. We compared the BMI trajectory between the offspring of mothers with and without GDM or type 1 diabetes, and the association between GDM and overweight offspring was analysed. Results: The prevalence of GDM and type 1 diabetes increased markedly over the study period. The BMI trajectory in the GDM or type 1 diabetes offspring differed significantly from the nondiabetic offspring. The timing of adiposity rebound occurred significantly earlier in the GDM (4.8 years) and type 1 diabetes (4.4 years) groups than the nondiabetic group (5.5 years). GDM offspring were more likely to be overweight at five, seven and 12 years of age (24.6%, 28.1%, 29.4%) than nondiabetic offspring (15.6%, 18.3%, 18.1%). Conclusion: Children born to mothers with GDM were significantly more likely to be overweight at an early age than those born to nondiabetic mothers. Abbreviations BMI, Body mass index; GDM, Gestational diabetes mellitus.
Työelämän tutkimus, Nov 1, 2007
arXiv (Cornell University), Sep 19, 2022
We investigate a novel non-parametric regression-based clustering algorithm for longitudinal data... more We investigate a novel non-parametric regression-based clustering algorithm for longitudinal data analysis. Combining natural cubic splines with Gaussian mixture models (GMM), the algorithm can produce smooth cluster means that describe the underlying data well. However, there are some shortcomings in the algorithm: high computational complexity in the parameter estimation procedure and a numerically unstable variance estimator. Therefore, to further increase the usability of the method, we incorporated approaches to reduce its computational complexity, we developed a new, more stable variance estimator, and we developed a new smoothing parameter estimation procedure. We show that the developed algorithm, SMIXS, performs better than GMM on a synthetic dataset in terms of clustering and regression performance. We demonstrate the impact of the computational speed-ups, which we formally prove in the new framework. Finally, we perform a case study by using SMIXS to cluster vertical atmospheric measurements to determine different weather regimes.
European journal of public health, Oct 20, 2017
an increase of 5.2 percentage points (95% CI 4.8-5.6); among males, the increase was 2.7 percenta... more an increase of 5.2 percentage points (95% CI 4.8-5.6); among males, the increase was 2.7 percentage points into 8.9% (95% CIs 2.4-2.9, 8.5%-9.4%). Adjusted hazard ratios of SSU, contrasting the cohorts, were 1.61 (95% CI 1.50-1.72) and 1.47 (95% CI 1.35-1.60) for females and males respectively. With the exception of schizophrenia, increased incidence was observed for all outcomes: hazard ratio estimates ranged from 1.15 to 6.11. Conclusions An increasing proportion of young people are coming to contact with specialized mental health related services during adolescence. Regardless of gender, this increase is evident across the varied spectrum of psychiatric and neurodevelopmental disorders and self-harm.
American Journal of Men's Health, Jul 14, 2009
http://ajmh.sagepub.com D espite the fact that fatherhood and father involvement in parenthood ha... more http://ajmh.sagepub.com D espite the fact that fatherhood and father involvement in parenthood have emerged as prominent areas of family research during the past decade (Lamb & Michael, 2000; Marsiglio, Amato, Day, & Lamb, 2000), the matter of father presence has attracted little attention in nursing and health care science. Social and behavioral scientists share the view that father's presence promotes the child's social,
Scandinavian Journal of Rheumatology, Aug 30, 2016
Objectives: To evaluate safety and effects of a new home treatment method, a whole-body cold mist... more Objectives: To evaluate safety and effects of a new home treatment method, a whole-body cold mist treatment, on chronic inflammatory arthritis patients. Method: Whole-body cold mist shower therapy was given to 121 voluntary patients with chronic inflammatory arthritis in this cross-over study during one week rehabilitation periods. Pain and sleep quality were assessed by a ten centimeter visual analogue scale (VAS). Mental status was assessed by DEPS-depression score. Body temperature, blood pressure, heart rate, use of occasional pain and sleep medication, and possible side effects were recorded. Results: The differences in pain (VAS) between treatment and control periods were significant (2.0 vs 2.4, p=0.006, paired t-test) in the last measurement, when assessed the pain of the passed week as a whole. A trend could be seen to an increasing difference towards the end of the week. The treatment effect was statistically significant (LRT, p<0.0001) after controlling for period and sequence effects. There was an indication for a better sleep quality (VAS) during the treatment period (2.3 vs 2.7, p=0.058 paired t-test) when assessing the passed week as a whole. The mean depression scale (DEPS) values showed no difference between the periods (5.5 vs 5.0, p= 0.1874 paired t-test, at start, and 4.5 vs 4.1 p=0.29 paired t-test, at the end). No significant side effects were recorded. Conclusions: The new whole-body cold treatment method may offer a safe option for pain selftreatment at home, but further study is needed to find out the clinical significance of the effect after a longer use.
European Journal of Public Health
Background Somatic symptoms among adolescents are common, yet little is known about long-term tra... more Background Somatic symptoms among adolescents are common, yet little is known about long-term trajectories of somatic symptoms and the factors in adolescence that shape them. We examined individual, family and school-based factors at age 16 as predictors of trajectories of somatic symptoms over 27 years. Methods Participants from the Northern Swedish Cohort (n = 1001) responded to questions about individual factors (e.g. health behaviours), family factors (e.g. contact with parents, social and material adversity) and school satisfaction at age 16; as well as 10 somatic symptoms at ages 16, 18, 21, 30 and 43. Teacher assessments at age 16 included overall ability at school and peer relations. Age 16 predictors of somatic symptom trajectory group membership were analysed using multinomial logistic regression. Results Poor contact with mother and poor school satisfaction were significant predictors of adverse symptom trajectories among both men and women. Low birth weight and low paren...
Data Science and SDGs, 2021
Transfer: European Review of Labour and Research, 2018
The connection between high-involvement management (HIM), entailing heavy employee involvement, a... more The connection between high-involvement management (HIM), entailing heavy employee involvement, and employee well-being is a controversial and widely discussed topic. Clarifying how job satisfaction and stress are connected to HIM and job control (the control employees have over their work), this study is based on data from two Finnish sources: an employer survey investigating the extent of HIM within an organisation, and employee assessments of job control, stress and job satisfaction. Logistic regression models were used as the study method. In contrast to previous Finnish studies, our findings show that HIM seems hardly to benefit employee well-being. Especially in the public sector, the correlation between extensive HIM and employee well-being turned out to be negative. However, HIM in the private sector was positively related to job satisfaction. As expected, a high level of job control was regularly associated with greater well-being.
Journal of Youth Studies, 2020
This study explores young Finns' transition pathways into education and work over a nine-year per... more This study explores young Finns' transition pathways into education and work over a nine-year period (from the age of 18 to 26), with a special focus on difficulties with labour market attachment. The data were derived from national registers comprising all men (29,383) and women (27,784) born in 1987. Their labour market attachment was investigated using trajectory analysis with four variables (employment, education, unemployment, social security allowances) from 2005 to 2013. We identified ten trajectory groups, of which seven in men and eight in women were categorised as 'mainstream paths', reflecting relatively smooth transitions. The other trajectory groups were considered to assume 'straying paths': 11 per cent of both men and women followed these paths. The size and nature of these paths differed between men and women: it seems that men more often have serious difficulties with labour market attachment than women do. Overall, the results of the multiple variable trajectory analyses suggest that cross-sectional inspections of the Finnish register data overestimate the proportion of NEETs, particularly among men.
One of the most popular methods for modelling survival analysis data is the ubiquitous and time-h... more One of the most popular methods for modelling survival analysis data is the ubiquitous and time-honored proportional hazards model of Cox (1972). Popularity notwithstanding, in several cases, however, the proportional hazards assumption is found to be violated. Thus, important model extensions have been developed in the intervening years. One such extension is the so-called Frailty model (see for example Collett, 2015), which is based on the utilization of random effects. Another extension arises from the application of finite mixtures in the context of time-to-event data. Although finite mixture modelling tools are used in many fields of science, they have been less well-developed and used in survival analysis. Thus, a key aim of this article is to provide an interesting application of mixture modeling in survival analysis and to discuss aspects arising in its application. In this paper, we apply these techniques using real data from the research register of the Finnish Centre for Pensions; using pension insurance mortality data, we use the basic Cox proportional hazards model by incorporating finite mixture modelling techniques. Additional comparisons with frailty models are also provided.
Social adversities in adolescence predict unfavourable trajectories of internalized mental health... more Social adversities in adolescence predict unfavourable trajectories of internalized mental health symptoms
Response to reviewers Reviewer 1:Anu Molarius Reviewer's report: TThe authors have sufficiently r... more Response to reviewers Reviewer 1:Anu Molarius Reviewer's report: TThe authors have sufficiently replied to almost all of the comments raised in the review and revised the manuscript accordingly. I have two further comments that do not necessarily require changes in the manucsript. Response 1:1. Thank you for this.
Data Science and SDGs, 2021
New Advances in Statistics and Data Science, 2017