In recent years, with the explosive adoption of smart phone devices, mobile health and fitness applications have been increasingly used by healthcare practitioners and the general public to manage electronic health records, chronic medical conditions, dietary references etc. Despite the rapid growth in the number of mobile and fitness applications on various platforms, very little work has been done to quantitatively and qualitatively assess these applications to guide users in the selection process. Automatic categorization of mobile health and fitness applications is the first step in this direction. In this paper, we report results from crawling 1,430 Android and 62,286 iOS apps in Nov. 2013. Among them, 1,399 apps were manually classified to one or multiple categories out of a total of 11 categories. Text mining tools were applied to the description section of the apps for keyword extraction, feature selection and automatic categorization. The classifiers we experimented with have comparable performance with Linear SVC achieving the highest precision, recall and f1 scores of 0.89, 0.79 and 0.88, respectively.
[1]
R. Tibshirani,et al.
Diagnosis of multiple cancer types by shrunken centroids of gene expression
,
2002,
Proceedings of the National Academy of Sciences of the United States of America.
[2]
Harry Zhang,et al.
The Optimality of Naive Bayes
,
2004,
FLAIRS.
[3]
R. Brereton,et al.
Support vector machines for classification and regression.
,
2010,
The Analyst.
[4]
Anika Ashok,et al.
Guidance for Industry by U.S. Department of Health and Human Services—Food and Drug Administration—Center for Biologics Evaluation and Research (CBER)—February 1999
,
2009
.
[5]
Gérard Dreyfus,et al.
Single-layer learning revisited: a stepwise procedure for building and training a neural network
,
1989,
NATO Neurocomputing.
[6]
P. Greenwood,et al.
A Guide to Chi-Squared Testing
,
1996
.