Modeling virtual organizations with Latent Dirichlet Allocation: A case for natural language processing

Este artigo tem o objetivo da avaliar a aplicação de dois métodos automáticos eficientes na extração de palavras-chave, usados pelas comunidades da Linguística de Corpus e do Processamento da Língua Natural para gerar palavras-chave de textos literários: o WordSmith Tools e o Latent Dirichlet Allocation (LDA). As duas ferramentas escolhidas para este trabalho têm suas especificidades e técnicas diferentes de extração, o que nos levou a uma análise orientada para a sua performance. Objetivamos entender, então, como cada método funciona e avaliar sua aplicação em textos literários. Para esse fim, usamos análise humana, com conhecimento do campo dos textos usados. O método LDA foi usado para extrair palavras-chave por meio de sua integração com o Portal Min@s: Corpora de Fala e Escrita, um sistema geral de processamento de corpora, concebido para diferentes pesquisas de Linguística de Corpus. Os resultados do experimento confirmam a eficácia do WordSmith Tools e do LDA na extração de palavras-chave de um corpus literário, além de apontar que é necessária a análise humana das listas em um estágio anterior aos experimentos para complementar a lista gerada automaticamente, cruzando os resultados do WordSmith Tools e do LDA. Também indicam que a intuição linguística do analista humano sobre as listas geradas separadamente pelos dois métodos usados neste estudo foi mais favorável ao uso da lista de palavras-chave do WordSmith Tools.

Download Full-text

Latent Dirichlet Allocation in predicting clinical trial terminations

BMC Medical Informatics and Decision Making ◽

10.1186/s12911-019-0973-y ◽

2019 ◽

Vol 19 (1) ◽

Author(s):

Simon Geletta ◽

Lendie Follett ◽

Marcia Laugerman

Keyword(s):

Clinical Trial ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Structured Data ◽

Unstructured Data ◽

Future Research ◽

Funding Agencies ◽

Dirichlet Allocation

Abstract Background This study used natural language processing (NLP) and machine learning (ML) techniques to identify reliable patterns from within research narrative documents to distinguish studies that complete successfully, from the ones that terminate. Recent research findings have reported that at least 10 % of all studies that are funded by major research funding agencies terminate without yielding useful results. Since it is well-known that scientific studies that receive funding from major funding agencies are carefully planned, and rigorously vetted through the peer-review process, it was somewhat daunting to us that study-terminations are this prevalent. Moreover, our review of the literature about study terminations suggested that the reasons for study terminations are not well understood. We therefore aimed to address that knowledge gap, by seeking to identify the factors that contribute to study failures. Method We used data from the clinicialTrials.gov repository, from which we extracted both structured data (study characteristics), and unstructured data (the narrative description of the studies). We applied natural language processing techniques to the unstructured data to quantify the risk of termination by identifying distinctive topics that are more frequently associated with trials that are terminated and trials that are completed. We used the Latent Dirichlet Allocation (LDA) technique to derive 25 “topics” with corresponding sets of probabilities, which we then used to predict study-termination by utilizing random forest modeling. We fit two distinct models – one using only structured data as predictors and another model with both structured data and the 25 text topics derived from the unstructured data. Results In this paper, we demonstrate the interpretive and predictive value of LDA as it relates to predicting clinical trial failure. The results also demonstrate that the combined modeling approach yields robust predictive probabilities in terms of both sensitivity and specificity, relative to a model that utilizes the structured data alone. Conclusions Our study demonstrated that the use of topic modeling using LDA significantly raises the utility of unstructured data in better predicating the completion vs. termination of studies. This study sets the direction for future research to evaluate the viability of the designs of health studies.

Download Full-text

Capturing and analyzing social representations. A first application of Natural Language Processing techniques to reader’s comments in COVID-19 news. Argentina, 2020

10.31235/osf.io/3pcdu ◽

2020 ◽

Author(s):

German Rosati ◽

Laia Domenech ◽

Adriana Chazarreta ◽

Tomas Maguire

Keyword(s):

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Social Representations ◽

Web Crawler ◽

First Approximation ◽

Processing Techniques ◽

Dirichlet Allocation

We present a first approximation to the quantification of social representations about the COVID-19, using news comments. A web crawler was developed for constructing the dataset of reader’s comments. We detect relevant topics in the dataset using Latent Dirichlet Allocation, and analyze its evolution during time. Finally, we show a first prototype to the prediction of the majority topics, using FastText.

Download Full-text

Latent Dirichlet Allocation in Predicting Clinical Trial Terminations

10.21203/rs.2.12904/v4 ◽

2019 ◽

Author(s):

Simon Geletta ◽

Lendie Follett ◽

Marcia R Laugerman

Keyword(s):

Clinical Trial ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Structured Data ◽

Unstructured Data ◽

Future Research ◽

Funding Agencies ◽

Dirichlet Allocation

Abstract This study used natural language processing (NLP) and machine learning (ML) techniques to identify reliable patterns from within research narrative documents to distinguish studies that complete successfully, from the ones that terminate. Recent research findings have reported that at least ten percent of all studies that are funded by major research funding agencies terminate without yielding useful results. Since it is well-known that scientific studies that receive funding from major funding agencies are carefully planned, and rigorously vetted through the peer-review process, it was somewhat daunting to us that study-terminations are this prevalent. Moreover, our review of the literature about study terminations suggested that the reasons for study terminations are not well understood. We therefore aimed to address that knowledge gap, by seeking to identify the factors that contribute to study failures.Method: We used data from the clinicialTrials.gov repository, from which we extracted both structured data (study characteristics), and unstructured data (the narrative description of the studies). We applied natural language processing techniques to the unstructured data to quantify the risk of termination by identifying distinctive topics that are more frequently associated with trials that are terminated and trials that are completed. We used the Latent Dirichlet Allocation (LDA) technique to derive 25 “topics” with corresponding sets of probabilities, which we then used to predict study-termination by utilizing random forest modeling. We fit two distinct models – one using only structured data as predictors and another model with both structured data and the 25 text topics derived from the unstructured data.Results: In this paper, we demonstrate the interpretive and predictive value of LDA as it relates to predicting clinical trial failure. The results also demonstrate that the combined modeling approach yields robust predictive probabilities in terms of both sensitivity and specificity, relative to a model that utilizes the structured data alone.Conclusions: Our study demonstrated that the use of topic modeling using LDA significantly raises the utility of unstructured data in better predicating the completion vs. termination of studies. This study sets the direction for future research to evaluate the viability of the designs of health studies.

Download Full-text

Klasifikasi Laporan Keluhan Pelayanan Publik Berdasarkan Instansi Menggunakan Metode LDA-SVM

Jurnal Teknologi Informasi dan Ilmu Komputer ◽

10.25126/jtiik.2021863768 ◽

2021 ◽

Vol 8 (6) ◽

pp. 1265

Author(s):

Muhammad Alkaff ◽

Andreyan Rizky Baskara ◽

Irham Maulani

Keyword(s):

Support Vector Machine ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Confusion Matrix ◽

Support Vector ◽

Topic Distribution ◽

The Government ◽

Dirichlet Allocation

Sebuah sistem layanan untuk menyampaikan aspirasi dan keluhan masyarakat terhadap layanan pemerintah Indonesia, bernama Lapor! Pemerintah sudah lama memanfaatkan sistem tersebut untuk menjawab permasalahan masyarakat Indonesia terkait permasalahan birokrasi. Namun, peningkatan volume laporan dan pemilahan laporan yang dilakukan oleh operator dengan membaca setiap keluhan yang masuk melalui sistem menyebabkan sering terjadi kesalahan dimana operator meneruskan laporan tersebut ke instansi yang salah. Oleh karena itu, diperlukan suatu solusi yang dapat menentukan konteks laporan secara otomatis dengan menggunakan teknik Natural Language Processing. Penelitian ini bertujuan untuk membangun klasifikasi laporan secara otomatis berdasarkan topik laporan yang ditujukan kepada instansi yang berwenang dengan menggabungkan metode Latent Dirichlet Allocation (LDA) dan Support Vector Machine (SVM). Proses pemodelan topik untuk setiap laporan dilakukan dengan menggunakan metode LDA. Metode ini mengekstrak laporan untuk menemukan pola tertentu dalam dokumen yang akan menghasilkan keluaran dalam nilai distribusi topik. Selanjutnya, proses klasifikasi untuk menentukan laporan agensi tujuan dilakukan dengan menggunakan SVM berdasarkan nilai topik yang diekstraksi dengan metode LDA. Performa model LDA-SVM diukur dengan menggunakan confusion matrix dengan menghitung nilai akurasi, presisi, recall, dan F1 Score. Hasil pengujian menggunakan teknik split train-test dengan skor 70:30 menunjukkan bahwa model menghasilkan kinerja yang baik dengan akurasi 79,85%, presisi 79,98%, recall 72,37%, dan Skor F1 74,67%. AbstractA service system to convey aspirations and complaints from the public against Indonesia's government services, named Lapor! The Government has used the Government for a long time to answer the problems of the Indonesian people related to bureaucratic problems. However, the increasing volume of reports and the sorting of reports carried out by operators by reading every complaint that comes through the system cause frequent errors where operators forward the reports to the wrong agencies. Therefore, we need a solution that can automatically determine the report's context using Natural Language Processing techniques. This study aims to build automatic report classifications based on report topics addressed to authorized agencies by combining Latent Dirichlet Allocation (LDA) and Support Vector Machine (SVM). The topic-modeling process for each report was carried out using the LDA method. This method extracts reports to find specific patterns in documents that will produce output in topic distribution values. Furthermore, the classification process to determine the report's destination agency carried out using the SVM based on the value of the topics extracted by the LDA method. The LDA-SVM model's performance is measured using a confusion matrix by calculating the value of accuracy, precision, recall, and F1 Score. The test results using the train-test split technique with a 70:30 show that the model produces good performance with 79.85% accuracy, 79.98% precision, 72.37% recall, and 74.67% F1 Score

Download Full-text

Latent Dirichlet Allocation in Predicting Clinical Trial Terminations

10.21203/rs.2.12904/v3 ◽

2019 ◽

Author(s):

Simon Geletta ◽

Lendie Follett ◽

Marcia R Laugerman

Keyword(s):

Clinical Trial ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Structured Data ◽

Unstructured Data ◽

Future Research ◽

Funding Agencies ◽

Dirichlet Allocation

Abstract Background: This study used natural language processing (NLP) and machine learning (ML) techniques to identify reliable patterns from within research narrative documents to distinguish studies that complete successfully, from the ones that terminate. Recent research findings have reported that at least ten percent of all studies that are funded by major research funding agencies terminate without yielding useful results. Since it is well-known that scientific studies that receive funding from major funding agencies are carefully planned, and rigorously vetted through the peer-review process, it was somewhat daunting to us that study-terminations are this prevalent. Moreover, our review of the literature about study terminations suggested that the reasons for study terminations are not well understood. We therefore aimed to address that knowledge gap, by seeking to identify the factors that contribute to study failures.Method: We used data from the clinicialTrials.gov repository, from which we extracted both structured data (study characteristics), and unstructured data (the narrative description of the studies). We applied natural language processing techniques to the unstructured data to quantify the risk of termination by identifying distinctive topics that are more frequently associated with trials that are terminated and trials that are completed. We used the Latent Dirichlet Allocation (LDA) technique to derive 25 “topics” with corresponding sets of probabilities, which we then used to predict study-termination by utilizing random forest modeling. We fit two distinct models – one using only structured data as predictors and another model with both structured data and the 25 text topics derived from the unstructured data.Results: In this paper, we demonstrate the interpretive and predictive value of LDA as it relates to predicting clinical trial failure. The results also demonstrate that the combined modeling approach yields robust predictive probabilities in terms of both sensitivity and specificity, relative to a model that utilizes the structured data alone.Conclusions: Our study demonstrated that the use of topic modeling using LDA significantly raises the utility of unstructured data in better predicating the completion vs. termination of studies. This study sets the direction for future research to evaluate the viability of the designs of health studies.

Download Full-text

Latent Dirichlet Allocation in Predicting Clinical Trial Failures

10.21203/rs.2.12904/v1 ◽

2019 ◽

Author(s):

Simon Geletta ◽

Lendie Follett ◽

Marcia R Laugerman

Keyword(s):

Clinical Trial ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Peer Review Process ◽

Structured Data ◽

Unstructured Data ◽

Funding Agencies ◽

Dirichlet Allocation

Abstract Background : This study used natural language processing (NLP) and machine learning (ML) techniques to identify reliable patterns from within research narrative documents to distinguish studies that complete successfully, from the ones that terminate. Recent research findings have reported that at least ten percent of all studies that are funded by major research funding agencies terminate without yielding useful results. Since it is well-known that scientific studies that receive funding from major funding agencies are carefully planned, and rigorously vetted through the peer-review process, it was somewhat daunting to us that study-terminations are this prevalent. Moreover, our review of the literature about study terminations suggested that the reasons for study terminations are not well understood. We therefore aimed to address that knowledge gap, by seeking to identify the factors that contribute to study failures. Method : We used data from the clinicialTrials.gov repository, from which we extracted both structured data (study characteristics), and unstructured data (the narrative description of the studies). We applied natural language processing techniques to the unstructured data to quantify the risk of termination by identifying distinctive topics that are more frequently associated with trials that are terminated and trials that are completed. We used the Latent Dirichlet Allocation (LDA) technique to derived 25 “topics” with corresponding sets of probabilities, which we then used to predict study-termination by utilizing random forest modeling. We fit two distinct models – one using only structured data as predictors and another model with both structured data and the 25 text topics derived from the unstructured data. Results : In this paper, we demonstrate the interpretive and predictive value of LDA as it relates to predicting clinical trial failure. The results also demonstrate that the combined modeling approach yields robust predictive probabilities in terms of both sensitivity and specificity, relative to a model that utilizes the structured data alone. Conclusions : Our study demonstrated that the use of topic modeling using LDA significantly raises the utility of unstructured data in better predicating the completion vs. termination of studies.

Download Full-text

Latent Dirichlet Allocation in Predicting Clinical Trial Terminations

10.21203/rs.2.12904/v2 ◽

2019 ◽

Author(s):

Simon Geletta ◽

Lendie Follett ◽

Marcia R Laugerman

Keyword(s):

Clinical Trial ◽

Natural Language Processing ◽

Natural Language ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Peer Review Process ◽

Structured Data ◽

Unstructured Data ◽

Funding Agencies ◽

Dirichlet Allocation

Abstract Background : This study used natural language processing (NLP) and machine learning (ML) techniques to identify reliable patterns from within research narrative documents to distinguish studies that complete successfully, from the ones that terminate. Recent research findings have reported that at least ten percent of all studies that are funded by major research funding agencies terminate without yielding useful results. Since it is well-known that scientific studies that receive funding from major funding agencies are carefully planned, and rigorously vetted through the peer-review process, it was somewhat daunting to us that study-terminations are this prevalent. Moreover, our review of the literature about study terminations suggested that the reasons for study terminations are not well understood. We therefore aimed to address that knowledge gap, by seeking to identify the factors that contribute to study failures. Method : We used data from the clinicialTrials.gov repository, from which we extracted both structured data (study characteristics), and unstructured data (the narrative description of the studies). We applied natural language processing techniques to the unstructured data to quantify the risk of termination by identifying distinctive topics that are more frequently associated with trials that are terminated and trials that are completed. We used the Latent Dirichlet Allocation (LDA) technique to derived 25 “topics” with corresponding sets of probabilities, which we then used to predict study-termination by utilizing random forest modeling. We fit two distinct models – one using only structured data as predictors and another model with both structured data and the 25 text topics derived from the unstructured data. Results : In this paper, we demonstrate the interpretive and predictive value of LDA as it relates to predicting clinical trial failure. The results also demonstrate that the combined modeling approach yields robust predictive probabilities in terms of both sensitivity and specificity, relative to a model that utilizes the structured data alone. Conclusions : Our study demonstrated that the use of topic modeling using LDA significantly raises the utility of unstructured data in better predicating the completion vs. termination of studies.

Download Full-text

Information Extraction from Electronic Medical Records using Natural Language Processing Techniques

Journal of Applied Sciences and Environmental Management ◽

10.4314/jasem.v24i6.13 ◽

2020 ◽

Vol 24 (6) ◽

pp. 1027-1033

Author(s):

E. Ogbuju ◽

G.N. Obunadike

Keyword(s):

Natural Language Processing ◽

Natural Language ◽

Medical Record ◽

Electronic Medical Record ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Medical Practitioners ◽

Industry Standard ◽

Processing Techniques ◽

Dirichlet Allocation

Patients share key information about their health with medical practitioners during clinic consultations. These key information may include their past medications and allergies, current situations/issues, and expectations. The healthcare professionals store this information in an Electronic Medical Record (EMR). EMRs have empowered research in healthcare; information hidden in them if harnessed properly through Natural Language Processing (NLP) can be used for disease registries, drug safety, epidemic surveillance, disease prediction, and treatment. This work illustrates the application of NLP techniques to design and implement a Key Information Retrieval System (KIRS framework) using the Latent Dirichlet Allocation algorithm. The cross-industry standard process for data mining methodology was applied in an experiment with an EMR dataset from PubMed todemonstrate the framework. The new system extracted the common problems (ailments) and prescriptions across the five (5) countries presented in the dataset. The system promises to assist health organizations in making informed decisions with the flood of key information data available in their domain. Keywords: Electronic Medical Record, BioNLP, Latent Dirichlet Allocation

Download Full-text

Understanding Communication in an Online Cancer Forum: Content Analysis Study

JMIR Cancer ◽

10.2196/29555 ◽

2021 ◽

Vol 7 (3) ◽

pp. e29555

Author(s):

Anietie Andy ◽

Uduak Andy

Keyword(s):

Natural Language Processing ◽

Natural Language ◽

Family Member ◽

Language Processing ◽

Latent Dirichlet Allocation ◽

Family Members ◽

Support Needs ◽

Online Forums ◽

Linguistic Inquiry ◽

Hospital Visits

Background Cancer affects individuals, their family members, and friends, and increasingly, some of these individuals are turning to online cancer forums to express their thoughts/feelings and seek support such as asking cancer-related questions. The thoughts/feelings expressed and the support needed from these online forums may differ depending on if (1) an individual has or had cancer or (2) an individual is a family member or friend of an individual who has or had cancer; the language used in posts in these forums may reflect these differences. Objective Using natural language processing methods, we aim to determine the differences in the support needs and concerns expressed in posts published on an online cancer forum by (1) users who self-declare to have or had cancer compared with (2) users who self-declare to be family members or friends of individuals with or that had cancer. Methods Using latent Dirichlet allocation (LDA), which is a natural language processing algorithm and Linguistic Inquiry and Word Count (LIWC), a psycholinguistic dictionary, we analyzed posts published on an online cancer forum with the aim to delineate the language features associated with users in these different groups. Results Users who self-declare to have or had cancer were more likely to post about LDA topics related to hospital visits (Cohen d=0.671) and use words associated with LIWC categories related to health (Cohen d=0.635) and anxiety (Cohen d=0.126). By contrast, users who declared to be family members or friends tend to post about LDA topics related to losing a family member (Cohen d=0.702) and LIWC categories focusing on the past (Cohen d=0.465) and death (Cohen d=0.181) were more associated with these users. Conclusions Using LDA and LIWC, we show that there are differences in the support needs and concerns expressed in posts published on an online cancer forum by users with cancer compared with family members or friends of those with cancer. Hence, responders to online cancer forums need to be cognizant of these differences in support needs and concerns and tailor their responses based on these findings.

Download Full-text