Multimodal Language Independent App Classification Using Images and Text

Singla, Kushal; Mukherjee, Nilanjana; Bose, Joy

doi:10.1007/978-3-319-91947-8_14

Cited by 4 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, since the number of reliable apps is low in rhinitis, this strategy may be developed when more apps become available. Another aspect of relevance is the customer perspective 54,55 . Our study did not retrieve any apps, although MASK‐air has been assessed twice for this aspect 30 …”

Section: Discussionmentioning

confidence: 99%

“…This attitude may have dangerous implications since patients may receive an incorrect diagnosis or management strategy. More in‐depth analysis, such as the American Psychiatric Association's app evaluation model, may be needed as a way to critically assess an app by considering accessibility, privacy and security, clinical foundation, engagement and interoperability 54 . However, since the number of reliable apps is low in rhinitis, this strategy may be developed when more apps become available.…”

Section: Discussionmentioning

confidence: 99%

“…Even though the set of text analysis conditions developed for this study has achieved a good specificity/sensitivity trade‐off, it can be improved by further refining the condition set. Another potential improvement is the use of additional data along with the app's title and description, such as the metadata inside the app's source code 54 . The automatic screening of apps addressing conditions other than AR (based on the same methods), and even developing a front‐end allowing physicians and patients to search for relevant apps for potentially any disease, will be a topic for future studies, taking into account the ever‐growing importance of mobile apps for the management of chronic diseases.…”

Section: Discussionmentioning

confidence: 99%

“…Another aspect of relevance is the customer perspective. 54,55 Our study did not retrieve any apps, although MASK-air has been assessed twice for this aspect. 30 Such an automatic screening tool has multiple applications.…”

Section: Main Findingsmentioning

confidence: 99%

See 3 more Smart Citations

Automatic market research of mobile health apps for the self‐management of allergic rhinitis

Antó¹,

Sousa‐Pinto

Czarlewski³

et al. 2022

Clin Experimental Allergy

View full text Add to dashboard Cite

Background: Only a small number of apps addressing allergic rhinitis (AR) patients have been evaluated. This makes their selection difficult. We aimed to introduce a new approach to market research for AR apps, based on the automatic screening of Apple App and Google Play stores.Methods: A JavaScript programme was devised for automatic app screening, and applied in a market assessment of AR self-management apps. We searched the Google Play and Apple App stores of three countries (USA, UK and Australia) with the following search terms: "hay fever", "hayfever", "asthma", "rhinitis", "allergic rhinitis". Apps were eligible if symptoms were evaluated. Results obtained with the automatic programme were compared to those of a blinded manual search. As an example, we used the search to assess apps that can be used to design a combined medication score for AR.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Main Findingsmentioning

confidence: 99%

See 2 more Smart Citations

Automatic market research of mobile health apps for the self‐management of allergic rhinitis

Antó¹,

Sousa‐Pinto

Czarlewski³

et al. 2022

Clin Experimental Allergy

View full text Add to dashboard Cite

show abstract

“…Zhu et al [18] enriched the textual information of mobile Apps by exploiting the additional Web knowledge from the Web search engine and combine all the enriched textual information into the Maximum Entropy model to train a mobile app classifier, while [10], [19] leveraged information publicly available from the online stores where the apps are marketed. To complement the existing text-based approach for app categorization, Singla et al [6] presented an app categorization system that uses object detection and recognition in images associated with apps to generate a more accurate categorization. Considering that the original category labels of the market are not fine-grained, Liu et al [7] developed a framework to label the apps with fine-grained categorical information.…”

Section: A Methods Of Using App's Metadatamentioning

confidence: 99%

DroidARA: Android Application Automatic Categorization Based on API Relationship Analysis

et al. 2019

View full text Add to dashboard Cite

An application (app) market with well-managed categorization will help users with app search and recommendation. Current categorization methods in app markets mainly rely on manual operation. Existing approaches for automatic Android app categorization suffer from low efficiency and low accuracy due to insufficient analysis of features, or inappropriate choice of features. This will mislead users to download unrelated apps and is not conducive to market stability maintenance. In this paper, we propose DroidARA, an efficient automatic categorization method for Android apps based on API relationships analysis. Considering that the app category can be characterized by API relationships which represent the combinations and links among APIs, we design a complete system to generate API call graphs, extract the API relationships information, transform them into feature vectors and train the classifier. Firstly, the API calls are obtained through static analysis to generate API call graphs that contain the relationships among APIs. A novel matrix structure as well as related algorithm are designed to extract the API relationships information from API call graphs. After that, the matrix is transformed into vector according to two feature selection methods we designed for strengthening the use of effective information in the API relationships. A convolutional neural network (CNN) model is then trained with labeled samples of such feature vectors. To validate the feasibility of DroidARA, we conduct several categorization experiments on 19949 real apps of Google Play. The results demonstrate that DroidARA can achieve an average 88.9% accuracy in categorizing the apps into 24 categories, which outperforms existing methods by 18.5%.INDEX TERMS Android, categorization, API relationship, static analysis.

show abstract

Automatic market research of mobile health apps for the self-management of allergic rhinitis (Preprint)

Bousquet¹

2021

Preprint

View full text Add to dashboard Cite

BACKGROUND Only a small number of apps addressing allergic rhinitis patients have been evaluated, which makes their selection difficult. OBJECTIVE We aimed to introduce a new approach to market research for allergic rhinitis apps based on the automatic screening of Apple App and Google Play stores. METHODS A JavaScript programme was devised for automatic app screening and applied in a market assessment of allergic rhinitis self-management apps. We searched the Google Play and Apple App stores of three countries (USA, UK and Australia) with the following search terms: hay fever, hayfever, asthma, rhinitis, allergic rhinitis. Apps were eligible if symptoms were evaluated. Results obtained with the automatic programme were compared to those of a blinded manual search. As an example, we used the search to assess apps that can be used to design a combined medication score for allergic rhinitis. RESULTS The automatic search programme identified 39 potentially eligible apps out of a total of 1,593 retrieved apps. Each of the 39 apps were individually checked, with 20 being classified as relevant. The manual search identified 19 relevant apps (out of 6,750 screened apps). Combining both methods, a total of 21 relevant apps were identified, pointing to a sensitivity of 95% and a specificity of 99% for the automatic method. Among these 21 apps, only two could be used for the combined symptom medication score for allergic rhinitis. CONCLUSIONS The programmed algorithm presented herein is able to continuously retrieve all relevant allergic rhinitis apps in the Apple App and Google Play stores, with high sensitivity and specificity. This approach has the potential to unveil the gaps and unmet needs of the apps developed so far.

show abstract

Multimodal Language Independent App Classification Using Images and Text

Cited by 4 publications

References 6 publications

Automatic market research of mobile health apps for the self‐management of allergic rhinitis

Automatic market research of mobile health apps for the self‐management of allergic rhinitis

DroidARA: Android Application Automatic Categorization Based on API Relationship Analysis

Automatic market research of mobile health apps for the self-management of allergic rhinitis (Preprint)

Contact Info

Product

Resources

About