Unobserved Data Research Articles

Learning Recommendation Algorithm Based on Improved BP Neural Network in Music Marketing Strategy.

The growth and popularity of streaming music have changed the way people consume music, and users can listen to online music anytime and anywhere. By integrating various recommendation algorithms/strategies (user profiling, collaborative filtering, content filtering, etc.), we capture users' interests and preferences and recommend the content of interest to them. To address the sparsity of behavioral data in digital music marketing, which leads to inadequate mining of user music preference features, a metric ranking learning recommendation algorithm with fused content representation is proposed. Relative partial order relations are constructed using observed and unobserved behavioral data to enable the model to be fully trained, while audio feature extraction submodels related to the recommendation task are constructed to further alleviate the data sparsity problem, and finally, the preference relationships between users and songs are mined through metric learning. Convolutional neural networks are used to extract the high-level semantic features of songs, and then the high-level semantic features of songs extracted from the previous layer are reformed into a session time sequence list according to the time sequence of user listening in order to build a bidirectional recurrent neural network model based on the attention mechanism so that it can reduce the influence of noisy data and learn the strong dependencies between songs.

Prediction of Missing Data in Rainfall Dataset by using Simple Statistical Method

Almost all of the data obtained from hydrological station contains missing data. Usually, this problem occurs due to equipment failures, maintenance work and human error. Incomplete dataset will reduce the ability of a statistical analysis and can cause a bias estimation due to systematic differences between observed and unobserved data. In this study, four simple statistical method such as Series Mean, Average Mean Top Bottom, Linear Interpolation and Nearest Neighbour were applied to predict the missing values in a rainfall dataset. An annual daily data for rainfall from nine selected monitoring station (from 2009 until 2018) were described using descriptive statistic. Then, the dataset were randomly simulated into 4 percentages of missing (5%, 10%, 15% and 20%) by using statistical package for social sciences software. The performance of this imputation methods were evaluated by using four performance indicators namely Mean Absolute Error, Root Mean Squared Error, Prediction Accuracy, and Index of Agreement. Overall, Linear Interpolation method was selected as the best imputation method to predict the missing data in the rainfall dataset.

Functional principal component analysis for longitudinal data with informative dropout.

In longitudinal studies, the values of biomarkers are often informatively missing due to dropout. The conventional functional principal component analysis typically disregards the missing information and simply treats the unobserved data points as missing completely at random. As a result, the estimation of the mean function and the covariance surface might be biased, resulting in a biased estimation of the functional principal components. We propose the informatively missing functional principal component analysis (imFunPCA), which is well suited for cases where the longitudinal trajectories are subject to informative missingness. Computation of the functional principal components in our approach is based on the likelihood of the data, where information of both the observed and missing data points are incorporated. We adopt a regression-based orthogonal approximation method to decompose the latent stochastic process based on a set of orthonormal empirical basis functions. Under the case of informative missingness, we show via simulation studies that the performance of our approach is superior to that of the conventional ones. We apply our method on a longitudinal dataset of kidney glomerular filtration rates for patients post renal transplantation.

DMFMDA: Prediction of Microbe-Disease Associations Based on Deep Matrix Factorization Using Bayesian Personalized Ranking.

Identifying the microbe-disease associations is conducive to understanding the pathogenesis of disease from the perspective of microbe. In this paper, we propose a deep matrix factorization prediction model (DMFMDA) based on deep neural network. First, the disease one-hot encoding is fed into neural network, which is transformed into a low-dimensional dense vector in implicit semantic space via embedding layer, and so is microbe. Then, matrix factorization is realized by neural network with embedding layer. Furthermore, our model synthesizes the non-linear modeling advantages of multi-layer perceptron based on the linear modeling advantages of matrix factorization. Finally, different from other methods using square error loss function, Bayesian Personalized Ranking optimizes the model from a ranking perspective to obtain the optimal model parameters, which makes full use of the unobserved data. Experiments show that DMFMDA reaches average AUCs of 0.9091 and 0.9103 in the framework of 5-fold cross validation and Leave-one-out cross validation, which is superior to three the-state-of-art methods. In case studies, 10, 9 and 9 out of top-10 candidate microbes are verified by recently published literature for asthma, inflammatory bowel disease and colon cancer, respectively. In conclusion, DMFMDA is successful application of deep learning in the prediction of microbe-disease association.

Genomic Prediction Enhanced Sparse Testing for Multi-environment Trials.

“Sparse testing” refers to reduced multi-environment breeding trials in which not all genotypes of interest are grown in each environment. Using genomic-enabled prediction and a model embracing genotype × environment interaction (GE), the non-observed genotype-in-environment combinations can be predicted. Consequently, the overall costs can be reduced and the testing capacities can be increased. The accuracy of predicting the unobserved data depends on different factors including (1) how many genotypes overlap between environments, (2) in how many environments each genotype is grown, and (3) which prediction method is used. In this research, we studied the predictive ability obtained when using a fixed number of plots and different sparse testing designs. The considered designs included the extreme cases of (1) no overlap of genotypes between environments, and (2) complete overlap of the genotypes between environments. In the latter case, the prediction set fully consists of genotypes that have not been tested at all. Moreover, we gradually go from one extreme to the other considering (3) intermediates between the two previous cases with varying numbers of different or non-overlapping (NO)/overlapping (O) genotypes. The empirical study is built upon two different maize hybrid data sets consisting of different genotypes crossed to two different testers (T1 and T2) and each data set was analyzed separately. For each set, phenotypic records on yield from three different environments are available. Three different prediction models were implemented, two main effects models (M1 and M2), and a model (M3) including GE. The results showed that the genome-based model including GE (M3) captured more phenotypic variation than the models that did not include this component. Also, M3 provided higher prediction accuracy than models M1 and M2 for the different allocation scenarios. Reducing the size of the calibration sets decreased the prediction accuracy under all allocation designs with M3 being the less affected model; however, using the genome-enabled models (i.e., M2 and M3) the predictive ability is recovered when more genotypes are tested across environments. Our results indicate that a substantial part of the testing resources can be saved when using genome-based models including GE for optimizing sparse testing designs.

Logistic principal component analysis via non-convex singular value thresholding

Multivariate binary data is becoming abundant in current biological research. Logistic principal component analysis (PCA) is one of the commonly used tools to explore the relationships inside a multivariate binary data set by exploiting the underlying low rank structure. We re-expressed the logistic PCA model based on the latent variable interpretation of the generalized linear model on binary data. The multivariate binary data set is assumed to be the sign observation of an unobserved quantitative data set, on which a low rank structure is assumed to exist. However, the standard logistic PCA model (using exact low rank constraint) is prone to overfitting, which could lead to the divergence of some estimated parameters towards infinity. We propose to fit a logistic PCA model through non-convex singular value thresholding to alleviate the overfitting issue. An efficient Majorization-Minimization algorithm is implemented to fit the model and a missing value based cross validation (CV) procedure is introduced for the model selection. Our experiments on realistic simulations of imbalanced binary data and low signal-to-noise ratio show that the CV error based model selection procedure is successful in selecting the proposed model. Furthermore, the selected model demonstrates superior performance in recovering the underlying low rank structure compared to models with convex nuclear norm penalty and exact low rank constraint. Finally, a binary copy number aberration data set is used to illustrate the proposed methodology in practice.

4489 Association between surgery with general anesthesia and cognitive decline in older adults: analysis using shared parameter models for informative dropout

OBJECTIVES/GOALS: Recent studies have assessed the association between surgery with general anesthesia and cognitive decline in longitudinal cohorts of older adults. Patients diagnosed with dementia more frequently drop out of these longitudinal studies or are unable to complete the test battery. We revisit this aim with focus on methods for informative dropout. METHODS/STUDY POPULATION: We use data from the Mayo Clinic Study of Aging (MCSA), a longitudinal epidemiological study of the prevalence, incidence, and risk factors for mild cognitive impairment (MCI) and dementia. Our primary outcome of interest was global cognitive z-score, assessed at study visits every 15 months. We implement linear mixed effects models to assess the association between post-enrollment exposure to surgery/anesthesia and subsequent cognitive decline trajectories. Demented patients more frequently drop out of MCSA, so, subjects with the worst cognitive outcomes are unobserved and missing data may be informative. Since this missingness may be missing not at random, we use shared parameter models to analyze continuous cognitive outcomes while jointly modeling time to dementia. RESULTS/ANTICIPATED RESULTS: A total 1948 subjects, non-demented at baseline, from the MCSA were included. Median age was 79, 51% of subjects were male, and 16% had MCI at enrollment. Among median follow-up of 4 study visits over median 5.4 years, 172 patients developed dementia and dropped out from further assessments of cognitive function. In adjusted linear mixed effects models, our data suggest post-enrollment exposure to surgery/anesthesia is associated with a decline in cognitive function over time (change in slope = −0.07 standard deviations of cognitive z-score per year, 95%CI = −0.08, −0.05, p<.001). After adjusting for informative dropout using shared parameter models, surgery/anesthesia is associated with greater cognitive decline (change in slope = −0.14 per year, 95%CI = −0.16, −0.12, p<.001). DISCUSSION/SIGNIFICANCE OF IMPACT: We revisited a prior analysis by our group with consideration of informative dropout. Subjects who dropout due to dementia may have different trajectories of cognitive decline compared to non-demented subjects. Shared parameter models estimate the association between surgery/anesthesia and cognitive decline accounting for informative dropout.

Uncertainty Quantification Accounting for Model Discrepancy Within a Random Effects Bayesian Framework

With the advent of integrated computational materials engineering, there is a drive to exchange statistical confidence in a design obtained from repeated experimentation with one developed through modeling and simulation. Since these models are often missing physics or include incomplete knowledge or simplifying assumptions into their mathematical construct, they may not capture the physical system or process adequately over the entire domain. This can lead to a systematic discrepancy, otherwise known as model misfit, between the model output and the system it represents over at least part of the domain. However, by accounting for this discrepancy in uncertainty analyses, reliable inference, and prediction on the model parameters and material behavior can be made despite the missing physics. The previous statement is contingent on two conditions: (1) the discrepancy is systematic, and (2) the structure of the discrepancy is well understood, which is required to minimize issues with non-identifiability among unknown model components. We illustrate these insights via a case study of inference and prediction in a phenomenological meso-scale VPSC crystal plasticity model, which does not contain physics describing the elastic regime of deformation. Inference is performed via a Bayesian approach enabled by posterior simulation. Posterior uncertainty about unknown model parameters takes into account observation error, uncertainty stemming from aleatoric sample-to-sample variability, and model form error. Posterior uncertainty in the unknown Voce hardening parameters is propagated to generate a distribution of the stress–strain response in both the elastic and plastic regimes. Additionally, posterior predictive distributions are simulated to establish uncertainty bounds for future unobserved data.

The generalized Lindemann melting coefficient

Lindemann developed the melting temperature theory over 100 years ago, known as the Lindemann criterion. Its main assumption is that melting occurs when the root-mean-square vibration amplitude of ions and atoms in crystals exceeds a critical fraction, η of the inter-atomic spacing in crystals. The Lindemann coefficient η is undefined and scientific papers report different η values for different elements. Here we present previously unobserved data trends pointing to the fact that the Lindemann coefficient could be linked to the periodic groups of the periodic table, having an exact value for each element belonging to a given periodic group. We report 12 distinctive Lindemann coefficient values corresponding to 12 groups of the periodic table containing solid elements with identifiable melting temperature. Using these vales, the recalculation of the melting temperatures indicates a good match to the experimental values for 39 elements, corresponding to 12 out of 15 periodic groups. This newly observed result opens up the possibility of further refining the Lindemann melting criterion by stimulating analytical studies of the Lindemann coefficient in the light of this newly discovered result.

Sensitivity analysis for clinical trials with missing continuous outcome data using controlled multiple imputation: A practical guide.

Missing data due to loss to follow-up or intercurrent events are unintended, but unfortunately inevitable in clinical trials. Since the true values of missing data are never known, it is necessary to assess the impact of untestable and unavoidable assumptions about any unobserved data in sensitivity analysis. This tutorial provides an overview of controlled multiple imputation (MI) techniques and a practical guide to their use for sensitivity analysis of trials with missing continuous outcome data. These include δ- and reference-based MI procedures. In δ-based imputation, an offset term, δ, is typically added to the expected value of the missing data to assess the impact of unobserved participants having a worse or better response than those observed. Reference-based imputation draws imputed values with some reference to observed data in other groups of the trial, typically in other treatment arms. We illustrate the accessibility of these methods using data from a pediatric eczema trial and a chronic headache trial and provide Stata code to facilitate adoption. We discuss issues surrounding the choice of δ in δ-based sensitivity analysis. We also review the debate on variance estimation within reference-based analysis and justify the use of Rubin's variance estimator in this setting, since as we further elaborate on within, it provides information anchored inference.

Likelihood inference and the role of initial conditions for the dynamic panel data model

Lancaster (2002) proposes an estimator for the dynamic panel data model with homoskedastic errors and zero initial conditions. In this paper, we show this estimator is invariant to orthogonal transformations, but is inefficient because it ignores additional information available in the data. The zero initial condition is trivially satisfied by subtracting initial observations from the data. We show that differencing out the data further erodes efficiency compared to drawing inference conditional on the first observations.Finally, we compare the conditional method with standard random effects approaches for unobserved data. Standard approaches implicitly rely on normal approximations, which may not be reliable when unobserved data is very skewed with some mass at zero values. For example, panel data on firms naturally depend on the first period in which the firm enters a new state. It seems unreasonable, then, to assume that the process determining unobserved data is known or stationary. We can instead make inference on structural parameters by conditioning on the initial observations.

Fast Adaptively Weighted Matrix Factorization for Recommendation with Implicit Feedback

Recommendation from implicit feedback is a highly challenging task due to the lack of the reliable observed negative data. A popular and effective approach for implicit recommendation is to treat unobserved data as negative but downweight their confidence. Naturally, how to assign confidence weights and how to handle the large number of the unobserved data are two key problems for implicit recommendation models. However, existing methods either pursuit fast learning by manually assigning simple confidence weights, which lacks flexibility and may create empirical bias in evaluating user's preference; or adaptively infer personalized confidence weights but suffer from low efficiency.To achieve both adaptive weights assignment and efficient model learning, we propose a fast adaptively weighted matrix factorization (FAWMF) based on variational auto-encoder. The personalized data confidence weights are adaptively assigned with a parameterized neural network (function) and the network can be inferred from the data. Further, to support fast and stable learning of FAWMF, a new specific batch-based learning algorithm fBGD has been developed, which trains on all feedback data but its complexity is linear to the number of observed data. Extensive experiments on real-world datasets demonstrate the superiority of the proposed FAWMF and its learning algorithm fBGD.

A variational autoencoder solution for road traffic forecasting systems: Missing data imputation, dimension reduction, model selection and anomaly detection

Efforts devoted to mitigate the effects of road traffic congestion have been conducted since 1970s. Nowadays, there is a need for prominent solutions capable of mining information from messy and multidimensional road traffic data sets with few modeling constraints. In that sense, we propose a unique and versatile model to address different major challenges of traffic forecasting in an unsupervised manner. We formulate the road traffic forecasting problem as a latent variable model, assuming that traffic data is not generated randomly but from a latent space with fewer dimensions containing the underlying characteristics of traffic. We solve the problem by proposing a variational autoencoder (VAE) model to learn how traffic data are generated and inferred, while validating it against three different real-world traffic data sets. Under this framework, we propose an online unsupervised imputation method for unobserved traffic data with missing values. Additionally, taking advantage of the low dimension latent space learned, we compress the traffic data before applying a prediction model obtaining improvements in the forecasting accuracy. Finally, given that the model not only learns useful forecasting features but also meaningful characteristics, we explore the latent space as a tool for model and data selection and traffic anomaly detection from the point of view of traffic modelers.

Statistical Modeling of Longitudinal Data with Non-ignorable Non-monotone Missingness with Semiparametric Bayesian and Machine Learning Components.

In longitudinal studies, outcomes are measured repeatedly over time and it is common that not all the patients will be measured throughout the study. For example patients can be lost to follow-up (monotone missingness) or miss one or more visits (non-monotone missingness); hence there are missing outcomes. In the longitudinal setting, we often assume the missingness is related to the unobserved data, which is non-ignorable. Pattern-mixture models (PMM) analyze the joint distribution of outcome and patterns of missingness in longitudinal data with non-ignorable nonmonotone missingness. Existing methods employ PMM and impute the unobserved outcomes using the distribution of observed outcomes, conditioned on missing patterns. We extend the existing methods using latent class analysis (LCA) and a shared-parameter PMM. The LCA groups patterns of missingness with similar features and the shared-parameter PMM allows a subset of parameters to be different between latent classes when fitting a model. We also propose a method for imputation using distribution of observed data conditioning on latent class. Our model improves existing methods by accommodating data with small sample size. In a simulation study our estimator had smaller mean squared error than existing methods. Our methodology is applied to data from a phase II clinical trial that studies quality of life of patients with prostate cancer receiving radiation therapy.

Fast Tensor Factorization for Large-Scale Context-Aware Recommendation from Implicit Feedback

This paper presents a fast Tensor Factorization (TF) algorithm for context-aware recommendation from implicit feedback. For such a recommendation problem, the observed data indicate the (positive) association between users and items in some given contexts. For better accuracy, it has been shown essential to include unobserved data that indicate the negative user-item-context associations. As such unobserved data greatly outnumber the observed ones, for efficiency existing algorithms usually use only a small part of the unobserved data for model training. We show in this paper that it is possible, and beneficial, to use all the unobserved data in training a TF based context-aware recommender system. This is achieved by two technical innovations. First, we scrutinize the matrix computation of the closed-form solution and accelerate the computation by memorizing the repetitive computation. Second, we further boost the generalization and scalability by dropping out some pairwise interactions when updating user, item or context factors to prevent overfitting and to reduce the training time. The resulting whole-data based learning algorithm, referred to as DropTF in the paper, is efficient and scale well. Our evaluation on two small benchmark datasets and a million-scale large dataset demonstrates improved accuracy over some existing algorithms for context-aware recommendation.

Item recommendation by predicting bipartite network embedding of user preference

With the development of e-commerce, various methodologies have studied to improve recommendation performance. Recently, many deep learning based network embedding approaches are applied to the recommendation domain. However, these approaches still have several limitations, such as the problem of data sparseness and the changing in user preference over time, which cannot be considered. In this paper, we propose a novel method for item recommendation based on network embedding. First, we apply a bipartite network embedding to address the data sparsity problem. Bipartite network embedding is a vector representation method that reflects explicit (i.e., observed data) and implicit relations (i.e., unobserved data). Bipartite network embedding methodology can address the data sparsity problem by using implicit relationship information from applying the random walk approach. Second, we predict future bipartite network embedding of user preference by adopting a Kalman filter to consider the changes in user preferences. We have conducted experiments to evaluate the effectiveness and performance of the proposed recommendation method. Through experimentation, the proposed recommendation method is validated as outperforming than the existing approaches including existing network embedding methods.

Missing data in longitudinal studies: Comparison of multiple imputation methods in a real clinical setting.

Missing data represent a challenge in longitudinal studies. The aim of the study is to compare the performance of the multivariate normal imputation and the fully conditional specification methods, using real data set with missing data partially completed 2 years later. The data used came from an ongoing randomized controlled trial with 5-year follow-up. At a certain time, we observed a number of patients with missing data and a number of patients whose data were unobserved because they were not yet eligible for a given follow-up. Both unobserved and missing data were imputed. The imputed unobserved data were compared with the corresponding real information obtained 2 years later. Both imputation methods showed similar performance on the accuracy measures and produced minimally biased estimates. Despite the large number of repeated measures with intermittent missing data and the non-normal multivariate distribution of data, both methods performed well and was not possible to determine which was better.

Travel Recommendation via Fusing Multi-Auxiliary Information into Matrix Factorization

As an e-commerce feature, the personalized recommendation is invariably highly-valued by both consumers and merchants. The e-tourism has become one of the hottest industries with the adoption of recommendation systems. Several lines of evidence have confirmed the travel-product recommendation is quite different from traditional recommendations. Travel products are usually browsed and purchased relatively infrequently compared with other traditional products (e.g., books and food), which gives rise to the extreme sparsity of travel data. Meanwhile, the choice of a suitable travel product is affected by an army of factors such as departure, destination, and financial and time budgets. To address these challenging problems, in this article, we propose a Probabilistic Matrix Factorization with Multi-Auxiliary Information (PMF-MAI) model in the context of the travel-product recommendation. In particular, PMF-MAI is able to fuse the probabilistic matrix factorization on the user-item interaction matrix with the linear regression on a suite of features constructed by the multiple auxiliary information. In order to fit the sparse data, PMF-MAI is built by a whole-data based learning approach that utilizes unobserved data to increase the coupling between probabilistic matrix factorization and linear regression. Extensive experiments are conducted on a real-world dataset provided by a large tourism e-commerce company. PMF-MAI shows an overwhelming superiority over all competitive baselines on the recommendation performance. Also, the importance of features is examined to reveal the crucial auxiliary information having a great impact on the adoption of travel products.

Soft Sensor Modeling for Unobserved Multimode Nonlinear Processes Based on Modified Kernel Partial Least Squares With Latent Factor Clustering

To cope with the soft sensor modeling of unobserved multimode nonlinear processes, this paper proposes a modified kernel partial least squares (KPLS) by integrating latent factor clustering (LFC), called LFC-KPLS. In the proposed method, the process data are first divided into several batches orderly, and then projected onto the latent space by using the nonlinear functional expansion technology. In the latent space, partial least squares method is applied to compute the regression coefficients between the input variables and output variable of each batch. These regression coefficients, called the latent factors, can describe the functional relationships in the unobserved multimode data. Therefore, the latent factors are used for mode clustering so that the process data with similar functional relations can be clustered in one mode together. For each mode, the nonlinear soft sensor is established based on KPLS. To assign the mode of the online query sample, a mode identification strategy based on Bayesian inference is designed for the soft sensor online prediction. Finally, two cases studies are adopted to validate the proposed method.

Semisupervised Deep State-Space Model for Plant Growth Modeling.

The optimal control of sugar content and its associated technology is important for producing high-quality crops more stably and efficiently. Model-based reinforcement learning (RL) indicates a desirable action depending on the type of situation based on trial-and-error calculations conducted by an environmental model. In this paper, we address plant growth modeling as an environmental model for the optimal control of sugar content. In the growth process, fruiting plants generate sugar depending on their state and evolve via various external stimuli; however, sugar content data are sparse because appropriate remote sensing technology is yet to be developed, and thus, sugar content is measured manually. We propose a semisupervised deep state-space model (SDSSM) where semisupervised learning is introduced into a sequential deep generative model. SDSSM achieves a high generalization performance by optimizing the parameters while inferring unobserved data and using training data efficiently, even if some categories of training data are sparse. We designed an appropriate model combined with model-based RL for the optimal control of sugar content using SDSSM for plant growth modeling. We evaluated the performance of SDSSM using tomato greenhouse cultivation data and applied cross-validation to the comparative evaluation method. The SDSSM was trained using approximately 500 sugar content data of appropriately inferred plant states and reduced the mean absolute error by approximately 38% compared with other supervised learning algorithms. The results demonstrate that SDSSM has good potential to estimate time-series sugar content variation and validate uncertainty for the optimal control of high-quality fruit cultivation using model-based RL.