Towards Enhancing Mobile App Reviews: A Structured Approach to

User Review Entry, Analysis and Veriﬁcation

Omar Haggag, John Grundy

and Rashina Hoda

HumaniSE Lab, Department of Software Systems and Cybersecurity, Faculty of IT, Monash University, Australia

Keywords:

Mobile Apps, User Reviews, Categorisation, ChatGPT, GPT-4, STGT, Tagging, Analysis, App Stores,

Transparency.

Abstract:

We propose an approach to address the shortcomings of current mobile app review systems on platforms

such as the Apple App Store and Google Play. Currently, these platforms lack review categorisation and au-

thentication of genuine user feedback, posing signiﬁcant barriers for app developers and users. We propose

an approach combining socio-technical grounded theory (STGT) and advanced natural language processing

(NLP) tools such as GPT-4 to analyse user reviews, providing deeper insights into app functionalities, prob-

lems, and ultimately, user satisfaction. An interactive UI prototype is presented to demonstrate the use of

structured, veriﬁed feedback. This includes a novel review submission process with categorisation/tagging

and a ”veriﬁed download” tag to ensure review authenticity. The goal of our approach is to enhance the app

ecosystem by assisting developers in prioritising improvements and enabling users to make informed choices,

encouraging a more robust and user-centric digital marketplace.

1 INTRODUCTION

In the dynamic and ever-expanding world of mobile

applications, user reviews stand as an essential com-

ponent in the digital ecosystem, linking app develop-

ers and users by providing transparent feedback on an

app’s performance, usability, and overall value (Vasa

et al., 2012; Haggag et al., 2021). These insights are

useful for developers, highlighting both strengths and

areas needing improvement, thereby assisting their

development strategies. For users, these reviews act

as a guide through millions of available mobile apps,

helping them in making informed decisions based

on the shared experiences of others (Palomba et al.,

2018). Not only do these reviews inﬂuence personal

download, purchase and usage choices, but they also

shape the evolution of apps to continually meet user

expectations through their updates (Genc-Nayebi and

Abran, 2017).

For app developers, user reviews are a signiﬁcant

feedback mechanism, revealing how their app per-

forms in real-world scenarios, which might not be de-

tected in testing environments (Li et al., 2018; Hag-

gag, 2022). These reviews can reveal issues, from

bugs to user experience problems. They also play a

https://orcid.org/0000-0003-4928-7076

signiﬁcant role in analysing the success of updates

and new features, inﬂuencing the app’s developmental

direction. On the user side, reviews are a signiﬁcant

resource for potential users, offering a more authen-

tic look than promotional materials. Current users,

through reviews, can share insights, contributing to

the app’s ongoing development and building commu-

nity (Palomba et al., 2017).

However, a major challenge with the current re-

view systems on platforms such as the Apple App

Store and Google Play is the absence of review cat-

egorisation (Li et al., 2018). This complicates de-

velopers’ ability to accurately and effectively analyse

and respond to feedback, especially with reviews of-

ten covering multiple issues, spelling and grammar

mistakes, and are sometimes submitted in different

languages. Another signiﬁcant concern is distinguish-

ing genuine user reviews from fake ones submitted by

people or generated by bots. Currently, there is no

deﬁnitive way to indicate or verify if a review is gen-

uine, which can skew the perception and reliability

of the feedback (Martens and Maalej, 2019; Haggag

et al., 2022a; Haggag et al., 2022b). Furthermore,

for paid or subscription-based mobile apps without a

system in place to conﬁrm if a user has made a pur-

chase or subscribed to services within the app, there’s

no guarantee that reviews reﬂect real customer expe-

598

Haggag, O., Grundy, J. and Hoda, R.

Towards Enhancing Mobile App Reviews: A Structured Approach to User Review Entry, Analysis and Veriﬁcation.

DOI: 10.5220/0012701000003687

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 19th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2024), pages 598-604

ISBN: 978-989-758-696-5; ISSN: 2184-4895

riences. This gap presents a high risk of fake reviews,

where individuals who have not actually bought or

engaged with the app can leave misleading feedback,

potentially affecting the app’s reputation and user de-

cisions.

We propose the use of socio-technical grounded

theory (STGT) to provide a structured and more com-

prehensive approach to analysing app reviews, offer-

ing insights into both the technological aspects of the

app and the social context of user interactions (Hoda,

2021). By applying STGT, researchers and app de-

velopers can better understand patterns and themes

in user reviews that go beyond simple functional-

ity issues, understanding how user sentiments evolve

with app updates or how social inﬂuences shape app

perception (Hoda, 2021; Hoda, 2023; Fazzini et al.,

2022). This analysis, empowered by text analysis

tools using natural language processing such as GPT-

4 combined with manual coding, can lead to a bet-

ter understanding of the user-app relationship. Also,

it enables developers to make user-centric enhance-

ments and align the app more closely with user needs

and preferences in this digitally interconnected world

(Sanderson, 2023).

We conducted a study to better understand (i) the

limitations of current reviewing mechanisms in app

stores; (ii) the challenges faced by app developers,

current and potential app users dealing with user re-

views in the current structure; and (iii) key areas for

improvement. The key contributions of this work in-

clude:

• analysis of how a structured, authentic review sys-

tem can assist potential users in making more in-

formed decisions, enriching their app selection and

usage experience;

• designing an interactive UI prototype as a proof of

concept highlighting the impact of organised, reli-

able user feedback on the app development cycle,

particularly in terms of addressing user-speciﬁc is-

sues and feature enhancement;

• creating a structured review submission method-

ology using categorisation/tagging, grounded in

STGT principles, to streamline and optimise the ex-

traction of useful feedback;

• introduction of ”Veriﬁed Download” and ”Veriﬁed

Purchase” tags to enhance the credibility and au-

thenticity of user reviews, ensuring that feedback is

sourced from real users; and

• developing a tool prototype to illustrate how novel

NLP tools such as GPT-4 combined with STGT

can greatly improve the user review submission and

analysis process.

2 MOTIVATION

Unstructured and Uncategorised Reviews: Fig-

ure 1 shows an example of the style of the current

user interface of app reviews in the App Store and

Google Play. The current mechanism for submit-

ting user reviews on major platforms like the Apple

App Store and Google Play has signiﬁcant limitations

for app developers, current users, and potential users

alike (Iacob and Harrison, 2013; Martens and Maalej,

2019). The absence of categorisation in the review

feedback system leads to a large, unstructured text of

user opinions and experiences. For app developers,

going through this large amount of data to extract ac-

tionable insights is a challenging task. Key issues and

popular feature requests can be hidden among less rel-

evant content, slowing down the response and resolu-

tion time and potentially leading to misguided prior-

ities(Ciurumelea et al., 2017). Moreover, the lack of

review structure often results in valuable feedback be-

ing overlooked or lost.

Figure 1: Current App Review UIs Lacking Categorisation.

App users looking to understand speciﬁc aspects

of an app, such as its performance, usability, or partic-

ular features, must navigate through a large amount of

general and irrelevant reviews (Vu et al., 2015). This

process is time-consuming and can be overwhelming,

worsening the overall experience and possibly lead-

ing to misinformed decision-making by the app de-

signers and developers. The credibility of these re-

views is another concern. With the prevalence of bot-

generated reviews and the difﬁculty in knowing au-

thentic user experiences from fake ones, users often

struggle to know the true quality and reliability of an

app (Caldeira et al., 2017; Martens and Maalej, 2019).

By categorising user review data into multiple

Towards Enhancing Mobile App Reviews: A Structured Approach to User Review Entry, Analysis and Veriﬁcation

599

different themes or aspects, such as usability, fea-

tures, bugs, and user interface (UI), a more struc-

tured framework for both submitting, identifying, and

analysing user feedback can be achieved. This would

enable developers to quickly identify and prioritise

areas that require attention, enhancing the efﬁciency

and effectiveness of the development process. Po-

tential app users could easily ﬁnd the information

most relevant to their interests or concerns, leading

to a more satisfying and informed app selection pro-

cess. Leveraging the latest NLP tools can also support

more accurately capturing, categorizing and interpret-

ing the various nature of user reviews.

Verifying Quality of Reviews: Existing app review

systems lack mechanisms to verify whether user re-

views originate from actual app downloads or con-

ﬁrmed purchases (Martens and Maalej, 2019). This

raises concerns about the authenticity of the feedback,

potentially enabling an increase in fake reviews or

bot-generated content. Introducing a ”Veriﬁed Down-

load” status and possibly a ”Veriﬁed Purchase” tag

alongside user reviews could signiﬁcantly mitigate

these issues. Such veriﬁcation processes would en-

sure that the feedback comes from users who have

genuinely downloaded and interacted with the app,

enhancing the credibility and value of the reviews for

both developers and other users.

Review Submission Process: Existing research on

user reviews in app stores predominantly focuses on

qualitative and quantitative analyses of the content of

the reviews, with minimal emphasis on enhancing the

review submission process itself (Ciurumelea et al.,

2017; Huebner et al., 2018; Li et al., 2017; Fu et al.,

2013; Alqahtani and Orji, 2020). This misses the cru-

cial aspect of how user reviews are collected and or-

ganised to ensure quality and timeliness. We want to

provide a structured and systematic approach to the

review capture and analysis process.

3 METHOD

We propose to leverage socio-technical grounded

theory (STGT) and the natural language processing

(NLP) capabilities of GPT-4 to aid in better categori-

sation of app reviews, and advocate for the use of

Veriﬁed Download and Purchase tags. We aim for

this method not just to enhance the clarity and rele-

vance of user reviews, but also to increase the overall

trustworthiness and usefulness of app reviews for all

stakeholders.

3.1 User Reviews Classiﬁcation Process

Using STGT and NLP

Our proposed user review categorisation methodol-

ogy for app stores is a two-phase system designed

to enhance the utility and relevance of user feedback.

The categorisation process outlined in Figure 2 rep-

resents a structured approach to managing user re-

views on an app store. In the initial submission phase

- step 1, users write and submit their reviews, which

are then preprocessed using NLP techniques to extract

key terms and sentiments in steps 1.1 and 1.2. Users

can further tag their reviews with hashtags to high-

light speciﬁc elements in step 1.3, after which the sys-

tem, informed by STGT-based analysis, suggests rel-

evant aspects for categorisation as in step 1.4. Users

then have the opportunity to verify or adjust these sug-

gestions before the review is added to the database, as

in step 1.5

Post-submission, reviews are tagged with ”Veri-

ﬁed Download” or ”Veriﬁed Purchase” to indicate au-

thenticity as in Step 2.1, and developers are notiﬁed

of the new classiﬁed feedback in step 2.2. The pub-

lished reviews in step 2.3 then become part of a con-

tinuous learning cycle, where the combined NLP sys-

tem and STGT framework evolve based on emerging

trends from new user feedback, ultimately allowing

both users and developers to ﬁlter and leverage re-

views more effectively in steps 2.4 and 2.5.

User Review Submission

NLP system performs

review text preprocessing

Feature Extraction using

NLP techniques

Initial Tagging

(Optional User-Driven Step)

A user starts writing a new user

reviewon the App Store

User can verify or adjust the suggested

aspects to ensure they accurately

represent their review & submit

Suggest related aspects by STGT to users

based on extracted features and sentiment

Users and Developers can filter

reviews with Aspects and Features

Reviews Extraction and Translation

Extract a Single User Review

Detect the Language of the

Extracted User Review using

Google API

Translate any language to

English

Add the Review to

the Database

Fetch the App by ID

Set a New IP Address after

extracting 100 reviews

5.1 million reviews for 278

mHealth are extracted and

translated to English

Review can be Marked with

"Verified Download" or

"Verified Purchase" Tags

Developer Notification

of the new review with

its classification

Review is published

on the App Store

NLP system uses new reviews

to fine-tune its models

STGT framework is updated to reflect

emerging features and aspects in user

feedback

Phase I: Submission process

Phase II: Post Review

Submission

Stage 1: Review Publishing

Stage 2: Continuous Learning

Review Added to

the Database

Step 1.1

Step 1

Step 1.2 Step 1.3

Step 1.5 Step 1.4

Step 2

Step 2.2Step 2.1 Step 2.3

Step 2.5Step 2.4

Figure 2: Proposed user reviews submission process.

The proposed classiﬁcation process is initiated the

moment a user starts writing a review. As they type,

our NLP algorithm using GPT-4 will analyse the con-

tent in real-time, suggesting relevant themes and cate-

gories based on STGT-informed underlying models.

ENASE 2024 - 19th International Conference on Evaluation of Novel Approaches to Software Engineering

600

These suggestions are derived from a set of com-

mon ’seed’ themes added into the model and previ-

ously identiﬁed themes within the user feedback pool,

such as usability, functionality, performance, and cus-

tomer service. This predictive categorisation will help

users classify their reviews with the most relevant as-

pects, improving the structure and searchability of

their feedback. However, our proposed system is

designed to empower users with the freedom to se-

lect their own tags or create new ones, ensuring that

the categorisation remains ﬂexible and user-driven.

When new user-deﬁned aspects are introduced, they

are fed back into the STGT framework, which is sig-

niﬁcant for capturing the evolving landscape of user

experiences and expectations.

The STGT approach guides this adaptive process

by providing a socio-technical lens, ensuring that both

social aspects (like UX and satisfaction) and technical

aspects (such as app functionality and bug reports) are

captured and reﬂected in the evolving classiﬁcation

model. The theoretical framework acts as a backbone

for understanding the complex interplay between the

app’s technical features and the social context of its

use. Speciﬁcs of its application will inevitably require

experimentation and revisions in practice.

Simultaneously, NLP algorithms work in the

background to reﬁne and expand the existing clas-

siﬁcation model. They process the natural language

of the reviews, adapting to new slang, terminologies,

and emerging issues. This dynamic approach guar-

antees that the classiﬁcation system remains up-to-

date with the latest trends and user concerns. Once

a review is submitted, it undergoes further analysis

to conﬁrm the initial categorisation. The NLP sys-

tem revisits the content, applying more rigorous text

analysis techniques to ensure that the ﬁnal categori-

sation aligns with the inductive analysis approach of

the STGT framework applied to this context. This in-

cludes correlating the user-selected aspects with the

model’s suggestions and adjusting the categorisation

model accordingly if needed.

Figure 3 illustrates the proposed interface in use,

designed using Figma. Visual Cues and an accessible

tagging system make it easy for users to contribute

to the classiﬁcation while expressing their feedback

in a structured manner that directly informs app de-

velopment and improves the UX for future app it-

erations. This will facilitate a clear, precise review

classiﬁcation process that is not only automated and

efﬁcient but also deeply rooted in the genuine needs

and contributions of the app’s user base. This dual

approach ensures that the review system evolves in

conjunction with user expectations and the app’s de-

velopment, leading to a robust mechanism for quality

feedback and continuous improvement.

Figure 3: Proposed User Interface of Review Submission –

see prototype here.

3.2 Implementation of ”Veriﬁed

Download” and ”Veriﬁed Purchase”

Tags

”Veriﬁed Download” and ”Veriﬁed Purchase” tags

alongside user reviews can improve the integrity and

relevance of feedback on app stores. For ”Veriﬁed

Download” , each time an app is downloaded, the app

store’s system records this event tied to the user’s ac-

count ID. When the user decides to leave a review for

the app, the system checks this record to conﬁrm if

a legitimate download has occurred. Upon successful

veriﬁcation, the user’s review is automatically marked

with a ”Veriﬁed Download” tag, as shown in Figure

4. This badge of authenticity is then displayed next

to the review on the app’s storefront to inform poten-

tial users that the feedback comes from an actual user

experience.

For ”Veriﬁed Purchase” , the system logs any in-

app purchases or subscriptions linked to the user’s ac-

count. When a review is posted, it cross-references

this data to verify whether the reviewer has made a

ﬁnancial commitment to the app. If a purchase or

subscription is veriﬁed, the review is marked with a

”Veriﬁed Purchase” tag as shown in Figure 4. This

process ensures that reviews reﬂecting the paid fea-

tures of the app are easily distinguishable, providing

prospective users with insights from those who have

fully engaged with the app’s offerings.

Technical implementation of these features will

need to prioritise data privacy, ensuring that only es-

sential data is used for veriﬁcation purposes and pro-

tecting against the falsiﬁcation of veriﬁcation tags.

The system should be capable of real-time updates,

reﬂecting the veriﬁed status immediately after a

Towards Enhancing Mobile App Reviews: A Structured Approach to User Review Entry, Analysis and Veriﬁcation

601

Figure 4: Reviews with veriﬁed download/purchase tags.

download or purchase occurs, regardless of the user’s

device or ﬁrmware version. Moreover, the design of

the ”Veriﬁed” tags within the user interface will be

clear to users but not disruptive to the overall UI of

the app store’s review section. By adopting these en-

hancements, app stores will offer a review system that

not only helps developers in obtaining genuine feed-

back but also enables users to recognise and trust the

authenticity of reviews, facilitating more informed de-

cisions regarding their app downloads and purchases.

3.3 Tool Prototype

We have developed a prototype tool to classify user

reviews and suggest relevant aspects in real-time as

the user types their feedback, as shown in Figure 5.

At the core of this system is a suite of NLP tech-

niques, primarily using a combination of Named En-

tity Recognition (NER) for extracting speciﬁc entities

and aspects from the text and Sentiment Analysis to

capture the emotional tone of the review. Leverag-

ing the power of Transformer-based models, partic-

ularly GPT-4, the tool dynamically processes the in-

put text to identify key themes and user sentiments.

Unlike Bidirectional Encoder Representations from

Transformers (BERT), which analyses text input bidi-

rectionally but independently of the context for each

word, GPT-4’s transformer architecture facilitates an

understanding of each word in relation to the en-

tire sentence structure, which signiﬁcantly enhances

the contextual relevance of aspect identiﬁcation and

sentiment interpretation. Concurrently, STGT analy-

sis framework is employed to interlink the extracted

entities and sentiments with broader socio-technical

pre-deﬁned aspects, providing users with intuitive as-

pect suggestions that reﬂect the context and content

of their reviews. This feature not only enhances the

review’s richness in detail but also aids in categoris-

ing the feedback for more actionable insights. The al-

gorithmic workﬂow is ﬁne-tuned through continuous

learning, using up-to-date user review data to reﬁne its

predictive capabilities and ensure high accuracy and

relevance in its suggestions.

Figure 5: A screenshot of our prototype tool for user review

classiﬁcation using GPT-4 and STGT.

4 EVALUATION OF BENEFITS

Enhanced Feedback Relevance and Prioritisation:

Developers beneﬁt from receiving feedback that is

both categorised and sentiment-analysed, which im-

proves the focus on user concerns that are most criti-

cal. This enables developers to prioritise updates and

features that will have the most signiﬁcant impact on

user satisfaction and app performance.

Quality Assurance and User Engagement Insights:

The incorporation of ”Veriﬁed Purchase” and ”Down-

load” tags assures developers that the feedback is

sourced from users who have genuinely interacted

with the app, providing a solid foundation for quality

assurance. Additionally, understanding user engage-

ment through the analysis of veriﬁed reviews informs

developers about which features or updates resonate

best with the users.

Resource Optimisation and Market Insight: The

clear categorisation of feedback streamlines the re-

view analysis process, allowing developers to allo-

cate their resources more effectively to address bugs

and develop new features. Furthermore, insights

from sentiment analysis and STGT offer a deeper un-

derstanding of market reception, which can inform

strategic business decisions.

Empowered User Feedback and Community

Building: For current app users, the visibility of cat-

egorised and valued feedback empowers them to pro-

vide more detailed input, knowing that their concerns

are recognised and acted upon. This not only en-

courages a richer dialogue but also fosters community

spirit as users witness their collective voice inﬂuenc-

ing app evolution.

Informed Decisions and Time Savings for Potential

Users: Potential app users gain the ability to make

more informed decisions based on the categorised and

veriﬁed reviews, which reﬂect real user experiences.

The categorisation of reviews by key aspects like us-

ability and performance means that potential users

can quickly ﬁnd the most signiﬁcant information, sav-

ENASE 2024 - 19th International Conference on Evaluation of Novel Approaches to Software Engineering

602

ing time and aiding in a more efﬁcient app selection

process.

Trust in App Quality and Risk Reduction: The

”Veriﬁed Purchase” tags signal a level of investment

and satisfaction from existing users, enhancing trust

in the app’s quality for potential users. This, coupled

with the insights from reviews of veriﬁed users, al-

lows potential users to assess the risks associated with

downloading or purchasing the app, ensuring they are

more comfortable and conﬁdent in their choices.

5 NEXT STEPS

In our future work, we aim to implement the integra-

tion of STGT with NLP techniques of our review clas-

siﬁcation system, particularly focusing on optimising

the precision of GPT-4 in sentiment and entity recog-

nition to better capture and analyse user feedback. A

signiﬁcant expansion would be the adaptation of the

system for direct integration of the classiﬁcation tool

with app development and feedback platforms, allow-

ing for a smooth feedback loop that could directly in-

ﬂuence app updates and feature enhancements. Ad-

ditionally, we plan to explore the application of pre-

dictive analytics to preemptively identify user trends

and enable proactive improvements to the app expe-

rience. This future work, prioritising algorithmic so-

phistication, cross-platform and multilingual support,

and predictive capabilities, is expected to signiﬁcantly

advance the responsiveness and user-centeredness of

app development practices.

ACKNOWLEDGEMENTS

Haggag and Grundy are supported by ARC Laureate

Fellowship FL190100035.

REFERENCES

Alqahtani, F. and Orji, R. (2020). Insights from user re-

views to improve mental health apps. Health infor-

matics journal, 26(3):2042–2066.

Caldeira, C., Chen, Y., Chan, L., Pham, V., Chen, Y., and

Zheng, K. (2017). Mobile apps for mood tracking:

an analysis of features and user reviews. In AMIA

Annual Symposium Proceedings, volume 2017, page

495. American Medical Informatics Association.

Ciurumelea, A., Schaufelb

uhl, A., Panichella, S., and Gall,

H. C. (2017). Analyzing reviews and code of mobile

apps for better release planning. In 2017 IEEE 24th

International Conference on Software Analysis, Evo-

lution and Reengineering (SANER), pages 91–102.

IEEE.

Fazzini, M., Khalajzadeh, H., Haggag, O., Li, Z., Obie, H.,

Arora, C., Hussain, W., and Grundy, J. (2022). Char-

acterizing human aspects in reviews of covid-19 apps.

Fu, B., Lin, J., Li, L., Faloutsos, C., Hong, J., and Sadeh, N.

(2013). Why people hate your app: Making sense of

user feedback in a mobile app store. In Proceedings of

the 19th ACM SIGKDD international conference on

Knowledge discovery and data mining, pages 1276–

1284.

Genc-Nayebi, N. and Abran, A. (2017). A systematic litera-

ture review: Opinion mining studies from mobile app

store user reviews. Journal of Systems and Software,

125:207–219.

Haggag, O. (2022). Better identifying and addressing di-

verse issues in mhealth and emerging apps using user

reviews. pages 329–335.

Haggag, O., Grundy, J., Abdelrazek, M., and Haggag, S.

(2022a). Better addressing diverse accessibility issues

in emerging apps: A case study using covid-19 apps.

Haggag, O., Grundy, J., Abdelrazek, M., and Haggag, S.

(2022b). A large scale analysis of mhealth app user

reviews. Empirical Software Engineering, 27(7):196.

Haggag, O., Haggag, S., Grundy, J., and Abdelrazek, M.

(2021). Covid-19 vs social media apps: Does privacy

really matter? pages 48–57.

Hoda, R. (2021). Socio-technical grounded theory for soft-

ware engineering. IEEE Transactions on Software En-

gineering, 48(10):3808–3832.

Hoda, R. (2023). Technical brieﬁng on socio-technical

grounded theory for qualitative data analysis. In 2023

IEEE/ACM 45th International Conference on Soft-

ware Engineering: Companion Proceedings (ICSE-

Companion), pages 344–345. IEEE.

Huebner, J., Frey, R. M., Ammendola, C., Fleisch, E., and

Ilic, A. (2018). What people like in mobile ﬁnance

apps: An analysis of user reviews. In Proceedings of

the 17th international conference on mobile and ubiq-

uitous multimedia, pages 293–304.

Iacob, C. and Harrison, R. (2013). Retrieving and analyz-

ing mobile apps feature requests from online reviews.

In 2013 10th working conference on mining software

repositories (MSR), pages 41–44. IEEE.

Li, X., Zhang, Z., and Stefanidis, K. (2018). Mobile app

evolution analysis based on user reviews. In New

Trends in Intelligent Software Methodologies, Tools

and Techniques, pages 773–786. IOS Press.

Li, Y., Jia, B., Guo, Y., and Chen, X. (2017). Mining user

reviews for mobile app comparisons. Proceedings of

the ACM on Interactive, Mobile, Wearable and Ubiq-

uitous Technologies, 1(3):1–15.

Martens, D. and Maalej, W. (2019). Towards understanding

and detecting fake reviews in app stores. Empirical

Software Engineering, 24(6):3316–3355.

Palomba, F., Linares-V

asquez, M., Bavota, G., Oliveto,

R., Di Penta, M., Poshyvanyk, D., and De Lucia, A.

(2018). Crowdsourcing user reviews to support the

evolution of mobile apps. Journal of Systems and Soft-

ware, 137:143–162.

Towards Enhancing Mobile App Reviews: A Structured Approach to User Review Entry, Analysis and Veriﬁcation

603

Palomba, F., Salza, P., Ciurumelea, A., Panichella, S., Gall,

H., Ferrucci, F., and De Lucia, A. (2017). Recom-

mending and localizing change requests for mobile

apps based on user reviews. In 2017 IEEE/ACM

39th International Conference on Software Engineer-

ing (ICSE), pages 106–117. IEEE.

Sanderson, K. (2023). Gpt-4 is here: what scientists think.

Nature, 615(7954):773.

Vasa, R., Hoon, L., Mouzakis, K., and Noguchi, A. (2012).

A preliminary analysis of mobile app user reviews. In

Proceedings of the 24th Australian computer-human

interaction conference, pages 241–244.

Vu, P. M., Pham, H. V., Nguyen, T. T., and Nguyen, T. T.

(2015). Tool support for analyzing mobile app re-

views. In 2015 30th IEEE/ACM International Con-

ference on Automated Software Engineering (ASE),

pages 789–794. IEEE.

ENASE 2024 - 19th International Conference on Evaluation of Novel Approaches to Software Engineering

604