Epileptic Focus Localization Based on iEEG by Using Positive Unlabeled (PU) Learning

C-PUGP: A cluster-based positive unlabeled learning method for disease gene prediction and prioritization

Tax evasion detection has a crucial role in addressing tax revenue loss. In the real world, an accessed tax dataset only contains a small number of labeled taxpayers who evade tax (positive samples) and a large number of unlabeled taxpayers who either evade tax or do not evade tax. It is difficult to address this issue due to this nontraditional dataset. In addition, the basic features of taxpayers designed according to tax experts’ domain knowledge and experience are very limited to determining whether taxpayers evade tax. These limitations motivate the contribution of this work. In this paper, we argue that the tax evasion detection task in the real world should be formalized as a positive unlabeled (PU) learning problem. We propose a novel tax evasion detection method based on PU learning with Network Embedding features (PUNE). PUNE effectively detects tax evasion based on basic features and transaction network features that are extracted by a network embedding algorithm. Moreover, PUNE can work well even under label noise. To evaluate the effectiveness of PUNE, we conduct experimental tests on a real-world tax dataset. The results demonstrate that PUNE can significantly improve the performance of tax evasion detection.

Conventional supervised binary classification algorithms have been widely applied to address significant research questions using biological and biomedical data. This classification scheme requires two fully labeled classes of data (e.g. positive and negative samples) to train a classification model. However, in many bioinformatics applications, labeling data is laborious, and the negative samples might be potentially mislabeled due to the limited sensitivity of the experimental equipment. The positive unlabeled (PU) learning scheme was therefore proposed to enable the classifier to learn directly from limited positive samples and a large number of unlabeled samples (i.e. a mixture of positive or negative samples). To date, several PU learning algorithms have been developed to address various biological questions, such as sequence identification, functional site characterization and interaction prediction. In this paper, we revisit a collection of 29 state-of-the-art PU learning bioinformatic applications to address various biological questions. Various important aspects are extensively discussed, including PU learning methodology, biological application, classifier design and evaluation strategy. We also comment on the existing issues of PU learning and offer our perspectives for the future development of PU learning applications. We anticipate that our work serves as an instrumental guideline for a better understanding of the PU learning framework in bioinformatics and further developing next-generation PU learning frameworks for critical biological applications.

Positive-unlabeled (PU) learning deals with the binary classification problem when only positive (P) and unlabeled (U) data are available, without negative (N) data. Existing PU methods perform well on the balanced dataset. However, in real applications such as financial fraud detection or medical diagnosis, data are always imbalanced. It remains unclear whether existing PU methods can perform well on imbalanced data. In this paper, we explore this problem and propose a general learning objective for PU learning targeting specially at imbalanced data. By this general learning objective, state-of-the-art PU methods based on optimizing a consistent risk can be adapted to conquer the imbalance. We theoretically show that in expectation, optimizing our learning objective is equivalent to learning a classifier on the oversampled balanced data with both P and N data available, and further provide an estimation error bound. Finally, experimental results validate the effectiveness of our proposal compared to state-of-the-art PU methods.

Positive-unlabeled learning for open set domain adaptation

Positive-unlabeled (PU) learning deals with the circumstances where only a portion of positive instances are labeled, while the rest and all negative instances are unlabeled, and due to this confusion, the class prior can not be directly available. Existing PU learning methods usually estimate the class prior by training a nontraditional probabilistic classifier, which is prone to give an overestimation. Moreover, these methods learn the decision boundary by optimizing the minimum margin, which is not suitable in PU learning due to its sensitivity to label noise. In this paper, we enhance PU learning methods from the above two aspects. More specifically, we first explicitly learn a transformation from unlabeled data to positive data by entropy regularized optimal transport to achieve a much more precise estimation for class prior. Then we switch to optimizing the margin distribution, rather than the minimum margin, to obtain a label noise insensitive classifier. Extensive empirical studies on both synthetic and real-world data sets demonstrate the superiority of our proposed method.

lncRNAs are involved in many biological processes, and their mutations and disorders are closely related to many diseases. Identification of LncRNA-Disease Associations (LDAs) helps us understand the pathogenesis of diseases and improve their diagnosis and treatment. However, experiments to determine LDAs are expensive, so it is essential to exploit effective computational methods to screen possible LDAs. In this study, we developed an LDA prediction model (LDA-DLPU) based on deep learning and positive-unlabeled (PU) learning. First, LDA-DLPU extracted features of lncRNAs and diseases based on singular value decomposition and regression model. Second, it selected negative LDAs based on PU learning and graph autoencoder. Finally, it classified unknown lncRNA-disease pairs based on deep neural network. LDA-DLPU obtained the best performance on two datasets. We predict that BCYRN1 and IFNG-AS1 may associate with leukemia.

In positive and unlabeled (PU) learning problems, only positive examples are labeled. Unlabeled data contain both positive and negative examples. Studies show that positive examples of (secondary) diagnoses, and clinical conditions, such as sepsis, are present in unlabeled hospital administrative data, potentially distorting hospital reimbursement systems, and negatively affecting hospitals' revenue and profitability. We investigate whether PU learning is suitable for improving the quality of hospital administrative data. We train three models on 313,434 hospital cases using hospital cost features: two based on the two-step "spy" approach and one using a robust PU learning method. For model evaluation, we rely exclusively on positive examples due to the PU setting. To further assess model performance, we perform an external validity check: We relabel unlabeled sepsis cases, derive new sepsis rates, and compare them to those reported in medical record review studies. All models identify true positives well in unseen data. External validity checks show, however, that only the robust PU learner effectively discriminates between positives and negatives in the unlabeled data, yielding new sepsis rates within the range of sepsis rates reported in medical record review studies. PU learning can improve the quality of hospital administrative data, but its effectiveness depends strongly on the choice of learning approach and classifier. The output of a PU learner can potentially improve hospital reimbursement systems, hospital revenue and profitability management, and sensitivity analyses in healthcare management science, health economics, health services research, and disease surveillance.

Successful cross-selling of products is a key goal of companies operating within the insurance industry. Choosing the right customer to approach for cross-purchase opportunities has a direct effect on both decreasing customer churn rate and increasing revenue. Unlike sales data of general products, insurance sales data typically contains only a few products (e.g., private medical insurance, life insurance, etc), it is highly imbalanced with a vast majority of customers with no cross-purchasing information, highly noisy due to varying purchase behaviour between different customers, and has no ground truth for knowing if the majority customers are truly non-cross-sell customers or they are missed opportunities. These data challenges render the building of machine learning models for accurately identifying potential cross-sell customers extremely difficult. This paper proposes a novel approach to solve this challenging problem of cross-sell customer identification using Positive Unlabelled (PU) learning in conjunction with advanced feature engineering on customer demographic data and unstructured customer question-response texts through topic modelling. We implement a bagging approach to iteratively learn the positive samples (the confirmed cross-sells) alongside random sub-samples of the unlabelled set. The proposed approach is extensively evaluated on real insurance data that has been newly collected from a leading insurance company for this study. Evaluation results demonstrate that our approach can successfully identify new potential opportunities for likely cross-sell customers.

A new approach for automatic detection of focal EEG signals using wavelet packet decomposition and quad binary pattern method

Deceptive reviews detection has attracted significant attention from both business and research communities. However, due to the difficulty of human labeling needed for supervised learning, the problem remains to be highly challenging. This paper proposed a novel angle to the problem by modeling PU (positive unlabeled) learning. A semi-supervised model, called mixing population and individual property PU learning (MPIPUL), is proposed. Firstly, some reliable negative examples are identified from the unlabeled dataset. Secondly, some representative positive examples and negative examples are generated based on LDA (Latent Dirichlet Allocation). Thirdly, for the remaining unlabeled examples (we call them spy examples), which can not be explicitly identified as positive and negative, two similarity weights are assigned, by which the probability of a spy example belonging to the positive class and the negative class are displayed. Finally, spy examples and their similarity weights are incorporated into SVM (Support Vector Machine) to build an accurate classifier. Experiments on gold-standard dataset demonstrate the effectiveness of MPIPUL which outperforms the state-of-the-art baselines.

A meta-learning framework using representation learning to predict drug-drug interaction

This study aims to develop an unsupervised automated method for detecting high-frequency oscillations (HFOs) in intracranial electroencephalogram (iEEG) signals, addressing the limitations of manual detection processes. The proposed method utilizes an unsupervised convolutional variational autoencoder (CVAE) model in conjunction with the short-term energy method (STE) to analyze two-dimensional time-frequency representations of iEEG signals. Candidate HFOs are identified using STE and transformed into time-frequency maps using the continuous wavelet transform (CWT). The CVAE model is trained for dimensionality reduction and feature reconstruction, followed by clustering of the reconstructed maps using the K-means algorithm for automated HFOs detection. Evaluation of the proposed unsupervised method on clinical iEEG data demonstrates its superior performance compared to traditional supervised models. The automated approach achieves an accuracy of 93.02%, sensitivity of 94.48%, and specificity of 92.06%, highlighting its efficacy in detecting HFOs with high accuracy. The unsupervised automated method developed in this study offers a reliable and efficient solution for detecting HFOs in iEEG signals, overcoming the limitations of manual detection processes of traditional supervised models. By providing clinicians with a clinically useful diagnostic tool, this approach holds promise for enhancing surgical resection planning in epilepsy patients and improving patient outcomes.

We introduce a new observational setting for Positive Unlabeled (PU) data where the observations at prediction time are also labeled. This occurs commonly in practice – we argue that the additional information is important for prediction, and call this task “augmented PU prediction”. We allow for labeling to be feature dependent. In such scenario, Bayes classifier and its risk is established and compared with a risk of a classifier which for unlabeled data is based only on predictors. We introduce several variants of the empirical Bayes rule in such scenario and investigate their performance. We put a special focus on dangers (and ease) of applying classical classification rule in the augmented PU scenario – due to no preexisting studies, an unaware researcher is prone to skewing the obtained predictions. We conclude that the variant based on recently proposed variational autoencoder designed for PU scenario works on par or better than other considered variants and yields advantage over feature-only based methods in terms of accuracy for unlabeled samples.

Intracranial electroencephalogram (iEEG) signals are generally recorded using multiple channels, and channel selection is therefore a significant means in studying iEEG-based seizure prediction. For n channels, [Formula: see text] channel cases can be generated for selection. However, by this means, an increase in n can cause an exponential increase in computational consumption, which may result in a failure of channel selection when n is too large. Hence, it is necessary to explore reasonable channel selection strategies under the premise of controlling computational consumption and ensuring high classification accuracy. Given this, we propose a novel method of channel reordering strategy combined with dual CNN-LSTM for effectively predicting seizures. First, for each patient with n channels, interictal and preictal iEEG samples from each single channel are input into the CNN-LSTM model for classification. Then, the F1-score of each single channel is calculated, and the channels are reordered in descending order according to the size of F1-scores (channel reordering strategy). Next, iEEG signals with an increasing number of channels are successively fed into the CNN-LSTM model for classification again. Finally, according to the classification results from n channel cases, the channel case with the highest classification rate is selected. Our method is evaluated on the three iEEG datasets: the Freiburg, the SWEC-ETHZ and the American Epilepsy Society Seizure Prediction Challenge (AES-SPC). At the event-based level, the sensitivities of 100%, 100% and 90.5%, and the false prediction rates (FPRs) of 0.10/h, 0/h and 0.47/h, are achieved for the three datasets, respectively. Moreover, compared to an unspecific random predictor, our method also shows a better performance for all patients and dogs from the three datasets. At the segment-based level, the sensitivities-specificities-accuracies-AUCs of 88.1%-94.0%-93.5%-0.9101, 99.1%-99.7%-99.6%-0.9935, and 69.2%-79.9%-78.2%-0.7373, are attained for the three datasets, respectively. Our method can effectively predict seizures and address the challenge of an excessive number of channels during channel selection.