The expansion of textual data, stemming from various sources such as online product reviews and scholarly publications on scientific discoveries, has created a significant demand for the extraction of succinct yet comprehensive information. While many methods have been proposed for automatic keyword extraction in unsupervised and fully supervised settings, effectively leveraging a partial list of known keywords, such as author-specified keywords or Twitter hashtags, remains under-explored. This work aims to enhance both the effectiveness and scalability of semi-supervised keyword extraction. We propose a novel variational Bayesian semi-supervised (VBSS) method that builds upon recent Bayesian advancement in the field, replacing computationally expensive posterior sampling with variational inference and data augmentation. This leads to closed-form updates and substantial speedups, particularly for long texts. Our numerical results show that the VBSS method not only improves performance on longer texts but also offers better control over false discovery rates compared to state-of-the-art keyword extraction techniques.

MOTIVATION: We propose a drug recommendation model that integrates information from both structured data (patient demographic information) and unstructured texts (patient reviews). It is based on multitask learning to predict review ratings of several satisfaction-related measures for a given medicine, where related tasks can learn from each other for prediction. The learned models can then be applied to new patients for drug recommendation. This is fundamentally different from most recommender systems in e-commerce, which do not work well for new customers (referred to as the cold-start problem). To extract information from review texts, we employ both topic modeling and sentiment analysis. We further incorporate variable selection into the model via Bayesian LASSO, which aims to filter out irrelevant features. To our best knowledge, this is the first Bayesian multitask learning method for ordinal responses. We are also the first to apply multitask learning to medicine recommendation. The sample code and data are made available at GitHub: https://github.com/thrushcyc-github/BMull. RESULTS: We evaluate the proposed method on two sets of drug reviews involving 17 depression/high blood pressure-related drugs. Overall, our method performs better than existing benchmark methods in terms of accuracy and AUC (area under the receiver operating characteristic curve). It is effective even with a small sample size and only a few available features, and more robust to possible noninformative covariates. Due to our model explainability, insights generated from our model may work as a useful reference for doctors. In practice, however, a final decision should be carefully made by combining the information from the proposed recommender with doctors' domain knowledge and past experience. AVAILABILITY AND IMPLEMENTATION: The sample code and data are publicly available at GitHub: https://github.com/thrushcyc-github/BMull.

In the era of big data, people benefit from the existence of tremendous amounts of information. However, availability of said information may pose great challenges. For instance, one big challenge is how to extract useful yet succinct information in an automated fashion. As one of the first few efforts, keyword extraction methods summarize an article by identifying a list of keywords. Many existing keyword extraction methods focus on the unsupervised setting, with all keywords assumed unknown. In reality, a (small) subset of the keywords may be available for a particular article. To use such information, we propose a rigorous probabilistic model based on a semisupervised setup. Our method incorporates the graph-based information of an article into a Bayesian framework via an informative prior so that our model facilitates formal statistical inference, which is often absent from existing methods. To overcome the difficulty arising from high-dimensional posterior sampling, we develop two Markov chain Monte Carlo algorithms based on Gibbs samplers and compare their performance using benchmark data. We use a false discovery rate (FDR)-based approach for selecting the number of keywords, whereas the existing methods use ad hoc threshold values. Our numerical results show that the proposed method compared favorably with state-of-the-art methods for keyword extraction. History: Accepted by Ramaswamy Ramesh, Area Editor for Data Science and Machine Learning. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2023.1283 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2021.0234 ) at ( http://dx.doi.org/10.5281/zenodo.7348935 ).

We propose a novel supervised dimension-reduction method called supervised t-distributed stochastic neighbor embedding (St-SNE) that achieves dimension reduction by preserving the similarities of data points in both feature and outcome spaces. The proposed method can be used for both prediction and visualization tasks with the ability to handle high-dimensional data. We show through a variety of data sets that when compared with a comprehensive list of existing methods, St-SNE has superior prediction performance in the ultrahigh-dimensional setting in which the number of features p exceeds the sample size n and has competitive performance in the p ≤ n setting. We also show that St-SNE is a competitive visualization tool that is capable of capturing within-cluster variations. In addition, we propose a penalized Kullback–Leibler divergence criterion to automatically select the reduced-dimension size k for St-SNE. Summary of Contribution: With the fast development of data collection and data processing technologies, high-dimensional data have now become ubiquitous. Examples of such data include those collected from environmental sensors, personal mobile devices, and wearable electronics. High-dimensionality poses great challenges for data analytics routines, both methodologically and computationally. Many machine learning algorithms may fail to work for ultrahigh-dimensional data, where the number of the features p is (much) larger than the sample size n. We propose a novel method for dimension reduction that can (i) aid the understanding of high-dimensional data through visualization and (ii) create a small set of good predictors, which is especially useful for prediction using ultrahigh-dimensional data.