Probabilistic and machine learning-based retrieval approaches for biomedical dataset retrieval

Abstract

The bioCADDIE dataset retrieval challenge brought together different approaches to retrieval of biomedical datasets relevant to a user’s query, expressed as a text description of a needed dataset. We describe experiments in applying a data-driven, machine learning-based approach to biomedical dataset retrieval as part of this challenge. We report on a series of experiments carried out to evaluate the performance of both probabilistic and machine learning-driven techniques from information retrieval, as applied to this challenge. Our experiments with probabilistic information retrieval methods, such as query term weight optimization, automatic query expansion and simulated user relevance feedback, demonstrate that automatically boosting the weights of important keywords in a verbose query is more effective than other methods. We also show that although there is a rich space of potential representations and features available in this domain, machine learning-based re-ranking models are not able to improve on probabilistic information retrieval techniques with the currently available training data. The models and algorithms presented in this paper can serve as a viable implementation of a search engine to provide access to biomedical datasets. The retrieval performance is expected to be further improved by using additional training data that is created by expert annotation, or gathered through usage logs, clicks and other processes during natural operation of the system. Database URL: https://github.com/emory-irlab/biocaddie

Discovering Related Terms and Detecting Trends in Software Engineering Using Word Embeddings

“…While messy, unstructured data makes up the majority of data created on a daily basis, estimated to be 95% of all "big data" [35]. Alongside IR's ability to read unstructured, semistructred data -data that possess attributes of both structured and unstructured -and structured data, it is the ability to consume the sheer quantity of data [36], the ability to filter data [37], as well as the ability to process and categorize data that has provided it's ultimate strength [31]. Categorizing data into various classes, the act of grouping alike data, can be accomplished in a variety of ways, one of which is clustering.…”

Data Curation and Optimization Techniques: A Systematic Mapping Study

“…This classification is available (in tabular form) at shorturl.at/D1234. [24], [15] [18], [17], [20], [22], [25], Techniques [27], [28], [43], [64], [30], [32], [34], [38], [40], [45], [46], [47], [49], [54], [55], [56], [57], [59], [60], [61], [66], [68], [69], [70], [73], [74], [75], [77], [26], [42], [44], [11], [13], [36]. [12], [14], [16], [19], [21], [23], [29], [31], [35], [39], [41], …”

A content-based dataset recommendation system for researchers—a case study on Gene Expression Omnibus (GEO) repository

“…In addition to that biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) dataset retrieval challenge was organized in 2016 to evaluate the effectiveness of information retrieval (IR) techniques in identifying relevant biomedical datasets in DataMed ( 3 ). Among the teams participated in this shared task, use of probabilistic or machine learning based IR ( 4 ), medical subject headings (MeSH) term based query expansion ( 5 ), word embeddings and identifying named entity ( 6 ), and re-ranking ( 7 ) for searching datasets using a query were the prevalent approaches. Similarly, a specialized search engine named Omicseq was developed for retrieving omics data ( 8 ).…”