Collecting sufficient labelled training data for health and medical problems is difficult (Antropova, et al., 2018). Also, missing values are unavoidable in health and medical datasets and tackling the problem arising from the inadequate instances and missingness is not straightforward (Snell, et al. 2017, Sterne, et al. 2009). However, machine learning algorithms have achieved significant success in many real-world healthcare problems, such as regression and classification and these techniques could possibly be a way to resolve the issues.
For constructing neuronal network models computational neuroscientists have access to wide-ranging anatomical data that nevertheless tend to cover only a fraction of the parameters to be determined. Finding and interpreting the most relevant data, estimating missing values, and combining the data and estimates from various sources into a coherent whole is a daunting task. With this chapter we aim to provide guidance to modelers by describing the main types of anatomical data that may be useful for informing neuronal network models. We further discuss aspects of the underlying experimental techniques relevant to the interpretation of the data, list particularly comprehensive data sets, and describe methods for filling in the gaps in the experimental data. Such methods of `predictive connectomics' estimate connectivity where the data are lacking based on statistical relationships with known quantities. It is instructive, and in certain cases necessary, to use organizational principles that link the plethora of data within a unifying framework where regularities of brain structure can be exploited to inform computational models. In addition, we touch upon the most prominent features of brain organization that are likely to influence predicted neuronal network dynamics, with a focus on the mammalian cerebral cortex. Given the still existing need for modelers to navigate a complex data landscape full of holes and stumbling blocks, it is vital that the field of neuroanatomy is moving toward increasingly systematic data collection, representation, and publication.
This paper investigates the sensitivity of forecast performance measures to taking a real time versus pseudo out-of-sample perspective. We use monthly vintages for the United States (US) and the Euro Area (EA) and estimate a set of vector autoregressive (VAR) models of different sizes with constant and time-varying parameters (TVPs) and stochastic volatility (SV). Our results suggest differences in the relative ordering of model performance for point and density forecasts depending on whether real time data or truncated final vintages in pseudo out-of-sample simulations are used for evaluating forecasts. No clearly superior specification for the US or the EA across variable types and forecast horizons can be identified, although larger models featuring TVPs appear to be affected the least by missing values and data revisions. We identify substantial differences in performance metrics with respect to whether forecasts are produced for the US or the EA.
This paper develops the inferential theory for latent factor models estimated from large dimensional panel data with missing observations. We propose an easy-to-use all-purpose estimator for a latent factor model by applying principal component analysis to an adjusted covariance matrix estimated from partially observed panel data. We derive the asymptotic distribution for the estimated factors, loadings and the imputed values under an approximate factor model and general missing patterns. The key application is to estimate counterfactual outcomes in causal inference from panel data. The unobserved control group is modeled as missing values, which are inferred from the latent factor model. The inferential theory for the imputed values allows us to test for individual treatment effects at any time under general adoption patterns where the units can be affected by unobserved factors.
Data from discovery proteomic and phosphoproteomic experiments typically include missing values that correspond to proteins that have not been identified in the analyzed sample. Replacing the missing values with random numbers, a process known as "imputation", avoids apparent infinite fold-change values. However, the procedure comes at a cost: Imputing a large number of missing values has the potential to significantly impact the results of the subsequent differential expression analysis. We propose a method that identifies differentially expressed proteins by ranking their observed changes with respect to the changes observed for other proteins. Missing values are taken into account by this method directly, without the need to impute them. We illustrate the performance of the new method on two distinct datasets and show that it is robust to missing values and, at the same time, provides results that are otherwise similar to those obtained with edgeR which is a state-of-art differential expression analysis method. The new method for the differential expression analysis of proteomic data is available as an easy to use Python package.
Background: Data preparation, such as missing values imputation and transformation, is the first step in any data analysis and requires crucial attention. Particularly, analysis of metabolites demands more preparation since those small compounds have recently been measurable in large scales with mass spectrometry techniques. We introduce novel statistical techniques for metabolite missing values imputation by utilizing replication samples. Results: To understand the nature of the missing values using replication samples, we obtained the empirical distribution of missing values and observed that the rate of missing values is approximately distributed as uniform across the metabolite range. Therefore, the missing values cannot be imputed with the lowest values. Using the identified distribution, we illustrated a simulation study to find an optimal imputation approach for metabolites. Conclusions: We demonstrated that the missing values in metabolomic data sets might not be necessarily low value. After identification of the nature of missing values, we validated K nearest neighborhood as an optimal approach for imputation.
The paper study recovery problem for discrete time signals with a finite number of missing values. The paper establishes recoverability of these missing values for signals with Z-transform vanishing with a certain rate at a single point. The transfer functions for the corresponding recovering kernels are presented explicitly. Some robustness of the recovery with respect to data truncation or noise contamination is established.
Due to the heterogeneity of the phenotype defined by Diagnostic and Statistical Manual of Mental Disorders (DSM) IV, it is not an optimal option to identify the genetic variation that underlies the risk for alcohol dependence (AD) and identifying subtypes of AD becomes an important topic. Traditional unsupervised cluster analysis and latent class analysis are the most commonly used methods to obtain the subtypes, but without the guidance of the genetic information, all these methods may lead to subtypes of little utility in genetic analysis. Recently, some multi-view co-clustering methods are proposed to ameliorate this drawback. However, these new methods did not take the missing values inside the data into consideration. To get around this limitation, we extended one of the multi-view methods to dealing with the missing values and clustering simultaneously. We applied this method to 2230 European-American sample and found that the well-known generic variant rs1229984 (in the ADH1B candidate gene) for the subtype is more significant than that corresponding to case-control association test. Finally, we verify it on the 1707 replication sample and find it significant, too.
Given sparse multi-dimensional data (e.g., (user, movie, time; rating) for movie recommendations), how can we discover latent concepts/relations and predict missing values? Tucker factorization has been widely used to solve such problems with multi-dimensional data, which are modeled as tensors. However, most Tucker factorization algorithms regard and estimate missing entries as zeros, which triggers a highly inaccurate decomposition. Moreover, few methods focusing on an accuracy exhibit limited scalability since they require huge memory and heavy computational costs while updating factor matrices. In this paper, we propose P-Tucker, a scalable Tucker factorization method for sparse tensors. P-Tucker performs alternating least squares with a row-wise update rule in a fully parallel way, which significantly reduces memory requirements for updating factor matrices. Furthermore, we offer two variants of P-Tucker: a caching algorithm P-Tucker-Cache and an approximation algorithm P-Tucker-Approx, both of which accelerate the update process. Experimental results show that P-Tucker exhibits 1.7-14.1x speed-up and 1.4-4.8x less error compared to the state-of-the-art. In addition, P-Tucker scales near linearly with the number of observable entries in a tensor and number of threads. Thanks to P-Tucker, we successfully discover hidden concepts and relations in a large-scale real-world tensor, while existing methods cannot reveal latent features due to their limited scalability or low accuracy.
Recommender system data presents unique challenges to the data mining, machine learning, and algorithms communities. The high missing data rate, in combination with the large scale and high dimensionality that is typical of recommender systems data, requires new tools and methods for efficient data analysis. Here, we address the challenge of evaluating similarity between two users in a recommender system, where for each user only a small set of ratings is available. We present a new similarity score, that we call LiRa, based on a statistical model of user similarity, for large-scale, discrete valued data with many missing values. We show that this score, based on a ratio of likelihoods, is more effective at identifying similar users than traditional similarity scores in user-based collaborative filtering, such as the Pearson correlation coefficient. We argue that our approach has significant potential to improve both accuracy and scalability in collaborative filtering.
An analysis of breast cancer incidences in women and the relationship between ethnicity and survival rate has been an ongoing study with recorded incidences of missing values in the secondary data. In this paper, we study and report the results of breast cancer survival rate by ethnicity, age and income groups from the dataset collected for 53593 patients in South East England between the years 1998 and 2003. In addition to this, we also predict the missing values for the ethnic groups in the dataset. The principle findings in our study suggest that: 1) women of white ethnicity in South East England have a highest percentage of survival rate when compared to the black ethnicity, 2) High income groups have higher survival rates to that of lower income groups and 3) Age groups between 80-95 have lower percentage of survival rate.
Plant survival is a key factor in forest dynamics and survival probabilities often vary across life stages. Studies specifically aimed at assessing tree survival are unusual and so data initially designed for other purposes often need to be used; such data are more likely to contain errors than data collected for this specific purpose. We investigate the survival rates of ten tree species in a dataset designed to monitor growth rates. As some individuals were not included in the census at some time points we use capture-mark-recapture methods both to allow us to account for missing individuals, and to estimate relocation probabilities. Growth rates, size, and light availability were included as covariates in the model predicting survival rates. The study demonstrates that tree mortality is best described as constant between years and size-dependent at early life stages and size independent at later life stages for most species of UK hardwood. We have demonstrated that even with a twenty-year dataset it is possible to discern variability both between individuals and between species. Our work illustrates the potential utility of the method applied here for calculating plant population dynamics parameters in time replicated datasets with small sample sizes and missing individuals without any loss of sample size, and including explanatory covariates.
Missing data is a common problem faced with real-world datasets. Imputation is a widely used technique to estimate the missing data. State-of-the-art imputation approaches, such as Generative Adversarial Imputation Nets (GAIN), model the distribution of observed data to approximate the missing values. Such an approach usually models a single distribution for the entire dataset, which overlooks the class-specific characteristics of the data. Class-specific characteristics are especially useful when there is a class imbalance. We propose a new method for imputing missing data based on its class-specific characteristics by adapting the popular Conditional Generative Adversarial Networks (CGAN). Our Conditional Generative Adversarial Imputation Network (CGAIN) imputes the missing data using class-specific distributions, which can produce the best estimates for the missing values. We tested our approach on benchmark datasets and achieved superior performance compared with the state-of-the-art and popular imputation approaches.
Multiple imputation (MI) is a popular method for dealing with missing values. However, the suitable way for applying clustering after MI remains unclear: how to pool partitions? How to assess the clustering instability when data are incomplete? By answering both questions, this paper proposed a complete view of clustering with missing data using MI. The problem of partitions pooling is here addressed using consensus clustering while, based on the bootstrap theory, we explain how to assess the instability related to observed and missing data. The new rules for pooling partitions and instability assessment are theoretically argued and extensively studied by simulation. Partitions pooling improves accuracy, while measuring instability with missing data enlarges the data analysis possibilities: it allows assessment of the dependence of the clustering to the imputation model, as well as a convenient way for choosing the number of clusters when data are incomplete, as illustrated on a real data set.
Time series are all around in real-world applications. However, unexpected accidents for example broken sensors or missing of the signals will cause missing values in time series, making the data hard to be utilized. It then does harm to the downstream applications such as traditional classification or regression, sequential data integration and forecasting tasks, thus raising the demand for data imputation. Currently, time series data imputation is a well-studied problem with different categories of methods. However, these works rarely take the temporal relations among the observations and treat the time series as normal structured data, losing the information from the time data. In recent, deep learning models have raised great attention. Time series methods based on deep learning have made progress with the usage of models like RNN, since it captures time information from data. In this paper, we mainly focus on time series imputation technique with deep learning methods, which recently made progress in this field. We will review and discuss their model architectures, their pros and cons as well as their effects to show the development of the time series imputation methods.
Clinical decision support using data mining techniques offers more intelligent way to reduce the decision error in the last few years. However, clinical datasets often suffer from high missingness, which adversely impacts the quality of modelling if handled improperly. Imputing missing values provides an opportunity to resolve the issue. Conventional imputation methods adopt simple statistical analysis, such as mean imputation or discarding missing cases, which have many limitations and thus degrade the performance of learning. This study examines a series of machine learning based imputation methods and suggests an efficient approach to in preparing a good quality breast cancer (BC) dataset, to find the relationship between BC treatment and chemotherapy-related amenorrhoea, where the performance is evaluated with the accuracy of the prediction.
This study introduces a framework for quality control of measured weather data, including anomaly detection, and infilling missing values. Weather data is a fundamental input to building performance simulations, in which anomalous values defect the results while missing data lead to an unexpected termination of the simulation process. Traditionally, infilling missing values in weather data is performed through periodic or linear interpolations. However, when missing values exceed many consecutive hours, the accuracy of traditional methods is subject to debate. This study demonstrates how Neural Networks can increase the accuracy of data imputation when compared to other supervised learning methods. The framework is validated by predicting missing temperature and relative humidity data for an observation site, through a network of nearby weather stations in Milan, Italy. Results show that the proposed method can facilitate real-time building simulations with accurate and rapid quality control.
Datasets with missing values are very common in real world applications. GAIN, a recently proposed deep generative model for missing data imputation, has been proved to outperform many state-of-the-art methods. But GAIN only uses a reconstruction loss in the generator to minimize the imputation error of the non-missing part, ignoring the potential category information which can reflect the relationship between samples. In this paper, we propose a novel unsupervised missing data imputation method named PC-GAIN, which utilizes potential category information to further enhance the imputation power. Specifically, we first propose a pre-training procedure to learn potential category information contained in a subset of low-missing-rate data. Then an auxiliary classifier is determined using the synthetic pseudo-labels. Further, this classifier is incorporated into the generative adversarial framework to help the generator to yield higher quality imputation results. The proposed method can improve the imputation quality of GAIN significantly. Experimental results on various benchmark datasets show that our method is also superior to other baseline approaches. Our code is available at \url{https://github.com/WYu-Feng/pc-gain}.
Supervised learning, characterized by both discriminative and generative learning, seeks to predict the values of single (or sometimes multiple) predefined target attributes based on a predefined set of predictor attributes. For applications where the information available and predictions to be made may vary from instance to instance, we propose the task of target-agnostic learning where arbitrary disjoint sets of attributes can be used for each of predictors and targets for each to-be-predicted instance. For this task, we survey a wide range of techniques available for handling missing values, self-supervised training and pseudo-likelihood training, and adapt them to a suite of algorithms that are suitable for the task. We conduct extensive experiments on this suite of algorithms on a large collection of categorical, continuous and discretized datasets, and report their performance in terms of both classification and regression errors. We also report the training and prediction time of these algorithms when handling large-scale datasets. Both generative and self-supervised learning models are shown to perform well at the task, although their characteristics towards the different types of data are quite different. Nevertheless, our derived theorem for the pseudo-likelihood theory also shows that they are related for inferring a joint distribution model based on the pseudo-likelihood training.
To study neurodegenerative diseases, longitudinal studies are carried on volunteer patients. During a time span of several months to several years, they go through regular medical visits to acquire data from different modalities, such as biological samples, cognitive tests, structural and functional imaging. These variables are heterogeneous but they all depend on the patient's health condition, meaning that there are possibly unknown relationships between all modalities. Some information may be specific to some modalities, others may be complementary, and others may be redundant. Some data may also be missing. In this work we present a neural network architecture for multimodal learning, able to use imaging and clinical data from two time points to predict the evolution of a neurodegenerative disease, and robust to missing values. Our multimodal network achieves 92.5\% accuracy and an AUC score of 0.978 over a test set of 57 subjects. We also show the superiority of the multimodal architecture, for up to 37.5\% of missing values in test set subjects' clinical measurements, compared to a model using only the clinical modality.