arXiv subjects

Yanshan Wang

Publications and source records attributed to Yanshan Wang.

61 records · Page 4Linked to original sources

arXiv2018

A Comparison of Word Embeddings for the Biomedical Natural Language Processing

Word embeddings have been widely used in biomedical Natural Language Processing (NLP) applications as they provide vector representations of words capturing the semantic properties of words and the linguistic relationship between words. Many biomedical applications use different textual resources (e.g., Wikipedia and biomedical articles) to train word embeddings and apply these word embeddings to downstream biomedical applications. However, there has been little work on evaluating the word embeddings trained from these resources.In this study, we provide an empirical evaluation of word embeddings trained from four different resources, namely clinical notes, biomedical publications, Wikipedia, and news. We performed the evaluation qualitatively and quantitatively. For the qualitative evaluation, we manually inspected five most similar medical words to a given set of target medical words, and then analyzed word embeddings through the visualization of those word embeddings. For the quantitative evaluation, we conducted both intrinsic and extrinsic evaluation. Based on the evaluation results, we can draw the following conclusions. First, the word embeddings trained on clinical notes and biomedical publications can capture the semantics of medical terms better, and find more relevant similar medical terms, and are closer to human experts' judgments, compared to these trained on Wikipedia and news. Second, there does not exist a consistent global ranking of word embedding quality for downstream biomedical NLP applications. However, adding word embeddings as extra features will improve results on most downstream tasks. Finally, the word embeddings trained on biomedical domain corpora do not necessarily have better performance than those trained on other general domain corpora for any downstream biomedical NLP tasks.

cs.IR↗

arXiv2018

A Deep Representation Empowered Distant Supervision Paradigm for Clinical Information Extraction

Objective: To automatically create large labeled training datasets and reduce the efforts of feature engineering for training accurate machine learning models for clinical information extraction. Materials and Methods: We propose a distant supervision paradigm empowered by deep representation for extracting information from clinical text. In this paradigm, the rule-based NLP algorithms are utilized to generate weak labels and create large training datasets automatically. Additionally, we use pre-trained word embeddings as deep representation to eliminate the need of task-specific feature engineering for machine learning. We evaluated the effectiveness of the proposed paradigm on two clinical information extraction tasks: smoking status extraction and proximal femur (hip) fracture extraction. We tested three prevalent machine learning models, namely, Convolutional Neural Networks (CNN), Support Vector Machine (SVM), and Random Forrest (RF). Results: The results indicate that CNN is the best fit to the proposed distant supervision paradigm. It outperforms the rule-based NLP algorithms given large datasets by capturing additional extraction patterns. We also verified the advantage of word embedding feature representation in the paradigm over term frequency-inverse document frequency (tf-idf) and topic modeling representations. Discussion: In the clinical domain, the limited amount of labeled data is always a bottleneck for applying machine learning. Additionally, the performance of machine learning approaches highly depends on task-specific feature engineering. The proposed paradigm could alleviate those problems by leveraging rule-based NLP algorithms to automatically assign weak labels and eliminating the need of task-specific feature engineering using word embedding feature representation.

cs.IR↗

arXiv2017

Generalized Ensemble Model for Document Ranking in Information Retrieval

A generalized ensemble model (gEnM) for document ranking is proposed in this paper. The gEnM linearly combines basis document retrieval models and tries to retrieve relevant documents at high positions. In order to obtain the optimal linear combination of multiple document retrieval models or rankers, an optimization program is formulated by directly maximizing the mean average precision. Both supervised and unsupervised learning algorithms are presented to solve this program. For the supervised scheme, two approaches are considered based on the data setting, namely batch and online setting. In the batch setting, we propose a revised Newton's algorithm, gEnM.BAT, by approximating the derivative and Hessian matrix. In the online setting, we advocate a stochastic gradient descent (SGD) based algorithm---gEnM.ON. As for the unsupervised scheme, an unsupervised ensemble model (UnsEnM) by iteratively co-learning from each constituent ranker is presented. Experimental study on benchmark data sets verifies the effectiveness of the proposed algorithms. Therefore, with appropriate algorithms, the gEnM is a viable option in diverse practical information retrieval applications.

cs.IR↗

arXiv2017

Stock price direction prediction by directly using prices data: an empirical study on the KOSPI and HSI

The prediction of a stock market direction may serve as an early recommendation system for short-term investors and as an early financial distress warning system for long-term shareholders. Many stock prediction studies focus on using macroeconomic indicators, such as CPI and GDP, to train the prediction model. However, daily data of the macroeconomic indicators are almost impossible to obtain. Thus, those methods are difficult to be employed in practice. In this paper, we propose a method that directly uses prices data to predict market index direction and stock price direction. An extensive empirical study of the proposed method is presented on the Korean Composite Stock Price Index (KOSPI) and Hang Seng Index (HSI), as well as the individual constituents included in the indices. The experimental results show notably high hit ratios in predicting the movements of the individual constituents in the KOSPI and HIS.

cs.CE↗

arXiv2016

MatLM: a Matrix Formulation for Probabilistic Language Models

Probabilistic language models are widely used in Information Retrieval (IR) to rank documents by the probability that they generate the query. However, the implementation of the probabilistic representations with programming languages that favor matrix calculations is challenging. In this paper, we utilize matrix representations to reformulate the probabilistic language models. The matrix representation is a superstructure for the probabilistic language models to organize the calculated probabilities and a potential formalism for standardization of language models and for further mathematical analysis. It facilitates implementations by matrix friendly programming languages. In this paper, we consider the matrix formulation of conventional language model with Dirichlet smoothing, and two language models based on Latent Dirichlet Allocation (LDA), i.e., LBDM and LDI. We release a Java software package--MatLM--implementing the proposed models. Code is available at: https://github.com/yanshanwang/JGibbLDA-v.1.0-MatLM.

cs.IR↗

arXiv2014

Indexing by Latent Dirichlet Allocation and Ensemble Model

The contribution of this paper is two-fold. First, we present Indexing by Latent Dirichlet Allocation (LDI), an automatic document indexing method. The probability distributions in LDI utilize those in Latent Dirichlet Allocation (LDA), a generative topic model that has been previously used in applications for document retrieval tasks. However, the ad hoc applications, or their variants with smoothing techniques as prompted by previous studies in LDA-based language modeling, result in unsatisfactory performance as the document representations do not accurately reflect concept space. To improve performance, we introduce a new definition of document probability vectors in the context of LDA and present a novel scheme for automatic document indexing based on LDA. Second, we propose an Ensemble Model (EnM) for document retrieval. The EnM combines basis indexing models by assigning different weights and attempts to uncover the optimal weights to maximize the Mean Average Precision (MAP). To solve the optimization problem, we propose an algorithm, EnM.B, which is derived based on the boosting method. The results of our computational experiments on benchmark data sets indicate that both the proposed approaches are viable options for document retrieval.

cs.IR↗

arXiv2013

A novel soft keyboard for touchscreen phones: QWERT

The popularity of touchscreen phones has been growing around the world since the iPhones and Android phones were released. More and more mobile phones with large touchscreen have been produced, however, the phones with small size displays are still in the majority of touch phones. The foremost interface on touch smartphones is the information input module using soft keyboards. Traditional input methods on touch phones have either too small key buttons (such as QWERTY) or too many functions (such as 3$\times$4 keyboard), which are inconvenient to use. Moreover, the conventional soft keyboards only use tapping to input texts while current touch smartphones allow various gestures on the touchscreen, such as sliding. In this paper, a novel soft keyboard called QWERT is proposed for touchscreen-based smartphones. The users can interact with phones via finger gestures of tapping or sliding when input text by using the QWERT. In doing so, the interactions between users and smartphones will be faster and easier. An experiment carried out on inexperienced human subjects shows that they can learn very fast due to their familiarities with QWERTY. A simulation experiment based on a cognitive architecture, ACT-R, was also conducted to predict the movement time (MT) of experienced human subjects. The simulation results show that the MT using QWERT outperforms other default keyboards. These outcomes imply that the novel QWERT is a viable option for touch smartphone users. Based on the novel design, an application is released on Android systems. This application is expected to give better user experience for customers who use touch smartphones.

cs.HC↗