Limits...
Mining Twitter Data to Improve Detection of Schizophrenia.

McManus K, Mallory EK, Goldfeder RL, Haynes WA, Tatum JD - AMIA Jt Summits Transl Sci Proc (2015)

Bottom Line: Using features from tweets such as emoticon use, posting time of day, and dictionary terms, we trained, built, and validated several machine learning models.Our support vector machine model achieved the best performance with 92% precision and 71% recall on the held-out test set.Additionally, we built a web application that dynamically displays summary statistics between cohorts.

View Article: PubMed Central - PubMed

Affiliation: Stanford University, Stanford, CA.

ABSTRACT
Individuals who suffer from schizophrenia comprise I percent of the United States population and are four times more likely to die of suicide than the general US population. Identification of at-risk individuals with schizophrenia is challenging when they do not seek treatment. Microblogging platforms allow users to share their thoughts and emotions with the world in short snippets of text. In this work, we leveraged the large corpus of Twitter posts and machine-learning methodologies to detect individuals with schizophrenia. Using features from tweets such as emoticon use, posting time of day, and dictionary terms, we trained, built, and validated several machine learning models. Our support vector machine model achieved the best performance with 92% precision and 71% recall on the held-out test set. Additionally, we built a web application that dynamically displays summary statistics between cohorts. This enables outreach to undiagnosed individuals, improved physician diagnoses, and destigmatization of schizophrenia.

No MeSH data available.


Related in: MedlinePlus

5-Fold cross validation results.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4525233&req=5

f2-2090726: 5-Fold cross validation results.

Mentions: SVM using the PCA transformed data performed the best in terms of F1 score and precision, and ANN using the PCA transformed features performing similarly, scoring slightly higher in recall [Figure 2]. It is worth noting that reducing the number of principal components used did not improve performance of the models. Given the number of features relative to the size of our data set, the models were able to fit the training data adequately with 28 features.


Mining Twitter Data to Improve Detection of Schizophrenia.

McManus K, Mallory EK, Goldfeder RL, Haynes WA, Tatum JD - AMIA Jt Summits Transl Sci Proc (2015)

5-Fold cross validation results.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4525233&req=5

f2-2090726: 5-Fold cross validation results.
Mentions: SVM using the PCA transformed data performed the best in terms of F1 score and precision, and ANN using the PCA transformed features performing similarly, scoring slightly higher in recall [Figure 2]. It is worth noting that reducing the number of principal components used did not improve performance of the models. Given the number of features relative to the size of our data set, the models were able to fit the training data adequately with 28 features.

Bottom Line: Using features from tweets such as emoticon use, posting time of day, and dictionary terms, we trained, built, and validated several machine learning models.Our support vector machine model achieved the best performance with 92% precision and 71% recall on the held-out test set.Additionally, we built a web application that dynamically displays summary statistics between cohorts.

View Article: PubMed Central - PubMed

Affiliation: Stanford University, Stanford, CA.

ABSTRACT
Individuals who suffer from schizophrenia comprise I percent of the United States population and are four times more likely to die of suicide than the general US population. Identification of at-risk individuals with schizophrenia is challenging when they do not seek treatment. Microblogging platforms allow users to share their thoughts and emotions with the world in short snippets of text. In this work, we leveraged the large corpus of Twitter posts and machine-learning methodologies to detect individuals with schizophrenia. Using features from tweets such as emoticon use, posting time of day, and dictionary terms, we trained, built, and validated several machine learning models. Our support vector machine model achieved the best performance with 92% precision and 71% recall on the held-out test set. Additionally, we built a web application that dynamically displays summary statistics between cohorts. This enables outreach to undiagnosed individuals, improved physician diagnoses, and destigmatization of schizophrenia.

No MeSH data available.


Related in: MedlinePlus