A supervised term ranking model for diversity enhanced biomedical information retrieval

Microbiome in Lower Urinary Tract Symptoms (LUTSs): Mapping the State of the Art with Bibliometric Analysis

BACKGROUND

The main objective of this study is to provide the first characterization of the current research field of the clinical microbiome in LUTSs.

METHODS

First-of-its-kind scientometric insight into the historical development and structural state of the discipline is provided by a field analysis, mapping, and sub-analysis of articles for future research. On 22 December 2022, the entire Scopus database was searched without language or date restrictions. Search terms included "Chronic prostatitis", OR "Interstitial cystitis", OR "Lower urinary tract symptoms", OR "Lower urinary tract dysfunction", OR "Overactive bladder", OR "Incontinence", OR "Urolithiasis", OR "Urothelium", OR "Urine", OR "Urology", OR "urinary disorder", OR "Pathophysiology", OR "Benign prostatic hyperplasia", OR "Benign prostatic enlargement", AND "Microbiota", OR "Microbiome", OR "Urobio-ma", OR "Urobiota; microflora". The author and institutional data were transformed using the analytical tool Biblioshiny (a Shiny app for Bibliometrix), which took into account variations in author spelling as well as institutional naming and subgroups.

RESULTS

The specified search strategy was able to locate 529 documents from 267 sources published from 1981 to 2022. The average number of years from publication was 4.59 years. The authors with the most publications were Wolfe AJ and Brubaker I. The top three most collaborative networks were Loyola University Chicago, Loyola University Medical Center, and the University of California San Diego. The most frequently occurring words among the 50 nodes were: human, humans, nonhuman, female, adult, article, microbiology, microflora, microbiota, and controlled study. Frontiers in Cellular and Infection Microbiology and the International Urogynecology Journal, followed by Nature Reviews Urology, were the top three most relevant sources in microbiome research in urology.

CONCLUSIONS

One of the most crucial requirements for developing research policies and anticipating the scientific requirements of researchers is paying attention to the evolution of various scientific fields. Understanding research gaps and future needs in microbiome research in urology can be effectively understood by paying attention to the models, maps, and visualizations used in this research, which are the results of systematic analysis of scientific products in the most esteemed scientific journals in the world.

Scientometric Study of Research in Information Retrieval in Medical Sciences

Background: Mapping scientific trends is one of the most important missions of scientometric research for effective research. The main goal of this paper was to visualize and draw the intellectual and cognitive structures of information retrieval (IR) in the medical sciences using science mapping.

Methods: In this cross-sectional scientometric study, we recruited all documents indexed in the Web of Science database with the topic of storing and retrieval of information in medical sciences. To analyze the results, 3 software, SciMAT-v1.1.04, VOSviewer-v1.6.14, CitNetExplorer_v1.0.0, were used.

Results: Our results showed that most scientific productions in this field fall into 2 categories: (1) effective methods of organizing information and (2) application and operation of the IR system in the process of intelligent questioning and answering, and analyzing information behaviors of physicians and health professionals. The results showed that the similarity index increased over time from 0.43 to 0.71. Analysis of the findings showed that similarity measures, expert systems, concepts, experience, answers, and multimodel IR clusters were considered as mature and completely centralized clusters in the first quarter of the strategic chart.

Conclusion: Because of the dramatic approximation of the vocabulary used by researchers and a relative slowdown in the growth rate of the subject's domain in the last decade, it seems necessary to pay attention to the expansion of the fields of IR and the application of its concepts in medical information sciences. Also, it can be recommended that designers of IR systems and techniques in medical information sciences pay more attention to human factors attentively to develop new technologies and tools.

A Query Expansion Method Using Multinomial Naive Bayes

Information retrieval (IR) aims to obtain relevant information according to a certain user need and involves a great diversity of data such as texts, images, or videos. Query expansion techniques, as part of information retrieval (IR), are used to obtain more items, particularly documents, that are relevant to the user requirements. The user initial query is reformulated, adding meaningful terms with similar significance. In this study, a supervised query expansion technique based on an innovative use of the Multinomial Naive Bayes to extract relevant terms from the first documents retrieved by the initial query is presented. The proposed method was evaluated using MAP and R-prec on the first 5, 10, 15, and 100 retrieved documents. The improved performance of the expanded queries increased the number of relevant retrieved documents in comparison to the baseline method. We achieved more accurate document retrieval results (MAP 0.335, R-prec 0.369, P5 0.579, P10 0.469, P15 0.393, P100 0.175) as compared to the top performers in TREC2017 Precision Medicine Track.

Integrated Random Negative Sampling and Uncertainty Sampling in Active Learning Improve Clinical Drug Safety Drug-Drug Interaction Information Retrieval

Clinical drug-drug interactions (DDIs) have been a major cause for not only medical error but also adverse drug events (ADEs). The published literature on DDI clinical toxicity continues to grow significantly, and high-performance DDI information retrieval (IR) text mining methods are in high demand. The effectiveness of IR and its machine learning (ML) algorithm depends on the availability of a large amount of training and validation data that have been manually reviewed and annotated. In this study, we investigated how active learning (AL) might improve ML performance in clinical safety DDI IR analysis. We recognized that a direct application of AL would not address several primary challenges in DDI IR from the literature. For instance, the vast majority of abstracts in PubMed will be negative, existing positive and negative labeled samples do not represent the general sample distributions, and potentially biased samples may arise during uncertainty sampling in an AL algorithm. Therefore, we developed several novel sampling and ML schemes to improve AL performance in DDI IR analysis. In particular, random negative sampling was added as a part of AL since it has no expanse in the manual data label. We also used two ML algorithms in an AL process to differentiate random negative samples from manually labeled negative samples, and updated both the training and validation samples during the AL process to avoid or reduce biased sampling. Two supervised ML algorithms, support vector machine (SVM) and logistic regression (LR), were used to investigate the consistency of our proposed AL algorithm. Because the ultimate goal of clinical safety DDI IR is to retrieve all DDI toxicity-relevant abstracts, a recall rate of 0.99 was set in developing the AL methods. When we used our newly proposed AL method with SVM, the precision in differentiating the positive samples from manually labeled negative samples improved from 0.45 in the first round to 0.83 in the second round, and the precision in differentiating the positive samples from random negative samples improved from 0.70 to 0.82 in the first and second rounds, respectively. When our proposed AL method was used with LR, the improvements in precision followed a similar trend. However, the other AL algorithms tested did not show improved precision largely because of biased samples caused by the uncertainty sampling or differences between training and validation data sets.