Source-linked AI summary
A Hybrid Usability Approach for Rating Evaluation of M-Commerce Applications
Ahmad Ibtisam, Bilal Khan, Arshad Ali
TL;DR
The study addresses limited attention to systematic prediction of M-commerce application ratings by examining usability factors. It develops a hybrid usability model and a stepwise regression model, finding a seven-predictor model with R2 = 0.910 and validation through K-fold analysis and PRED(25).
Problem
Systematic prediction of M-commerce application ratings has received limited research attention, motivating investigation of usability factors that determine ratings.
Method
The study combines a hybrid usability model with a questionnaire-based assessment and stepwise multiple linear regression to predict ratings from usability factors and criteria.
Results
A seven-predictor MLR model achieved R2 = 0.910, while fixed K-fold models showed significant results with good R2 and PRED(25) values equivalent to one.
Takeaways & Limitations
The paper provides a usability factor-based model and a regression-based approach for determining and predicting M-commerce application ratings.
Abstract
from arXiv · showhide
The success of any mobile application relies on its usefulness and rating is considered as an important measure in this regard. This research work focuses on identifying usability factors, which contribute significantly towards the rating of M-commerce apps. This work intends to explore existing usability models consisting of different factors along with a set of criteria and evaluate in terms of rating estimation by considering 5 well-known mobile applications, namely (i) daraz, (ii) shophive, (iii) home shopping, (iv) Symbios. (v) yayvo. Then, this work provides a hybrid usability model for rating prediction of M-commerce applications. The initial hybrid usability model comprises of (i) learnability, (ii) consistency, (iii) human factors,(iv)communicativeness,(v)effectiveness, (vi) Operability, (vii) efficiency, (viii) satisfaction. Each factor consists of some criteria. Keeping in view the factors of hybrid usability model, the data was collected from 40 users for each application. Furthermore, Forward Stepwise Multiple Linear Regression based rating prediction model is suggested by analyzing each criterion of all factors of hybrid usability model. Finally, the model is assessed and validated by using PRED(x) and K-fold techniques.
I. INTRODUCTION
Mobile commerce extends electronic commerce through internet-enabled mobile devices, making shopping more accessible while introducing usability and user-experience challenges. The study therefore examines usability factors associated with M-commerce application ratings.
- Mobile commerce requires transaction initiation and completion through mobile access on an internet-enabled device.
- M-commerce offers real-time information access, personalization, communication functionality, and continual availability compared with conventional commerce.
- Ease of use is a critical development concern because mobile applications create new user-experience challenges.
- The study investigates usability factors for determining M-commerce application ratings because systematic rating prediction has received limited research attention.
B. OBJECTIVES
The study aims to investigate usability factors affecting M-commerce application ratings and develop hybrid and regression-based models for rating determination and prediction.
- The research investigates the impact of various usability factors on M-commerce application ratings.
- The study explores existing usability models and their constituent factors.
- The proposed hybrid usability model is intended to determine ratings for various M-commerce applications.
- The study develops a regression-based model to predict ratings of Android-based M-commerce applications from usability factors.
C. RESEARCH METHODOLOGY
The research uses a quantitative survey-based process to identify usability attributes, estimate ratings with regression, and validate the resulting model.
- Existing usability models were studied before primary data were collected through a structured questionnaire based on prior instruments.
- The identified attributes were analyzed using linear and multiple regression for rating estimation.
- The model was subsequently assessed and validated after regression analysis.
- The paper organizes its remaining sections around literature review, Delphi-based data collection, prediction modeling, validation, and conclusion.
II. LITERATURE REVIEW
The literature review situates usability within established software-quality models and describes the questionnaire and usability concept used for the study.
- Software-quality models developed over three decades commonly treat usability as a major quality attribute.
- McCall’s model includes usability alongside correction, efficiency, integrity, and reliability.
- FURPS includes usability, reliability, performance, and compatibility as non-functional requirements.
- The study frames usability as the degree to which a product or system enables people with varied qualities and abilities to accomplish a specified objective in a defined context.
A. RELATED WORK
Prior studies examined usability through cultural qualities, simplicity, interaction, operating systems, and software-quality models. These studies informed research on mobile and M-commerce application usability.
- Prior usability studies: Cultural qualities were examined as moderators of social-media application usability across 1,844 customers from four countries.The qualities included masculinity, independence, power distance, uncertainty avoidance, and long-term orientation.
- Prior usability studies: Mobile application usability was studied in relation to simplicity and interaction among 310 Korean mobile clients.Partial Least Squares was used, with reliability above 0.80 and AVE above 0.50 for both variables.
- Prior usability studies: Usability was also experimentally assessed across iOS, Android, and Symbian mobile operating systems.The study developed a model using multiple measures to compare usability across operating systems.
- Prior usability studies: Application design was identified as a major issue affecting mobile-application usability, while the reported observations were limited.The passage connects usability and application design but does not specify the limits in greater detail.
III. DELPHI PROCESS
The study began with a 60-item, five-point Likert questionnaire covering 13 usability attributes. A Delphi process then finalized 40 questions for assessing M-commerce applications.
- Questionnaire design: The initial questionnaire contained 60 items rated on a five-point Likert scale across 13 usability attributes.Attributes included efficiency, learnability, satisfaction, human factors, communicativeness, effectiveness, operability, protection, consistency, accessibility, user-errors protection, memorability, and recognisability.
A. DATA COLLECTION
The study collected task-based usability assessments with a finalized questionnaire and used stepwise regression to model ratings from usability predictors. K-fold and MRE/MMRE procedures were used for assessment and validation.
- Data collection: Participants assessed M-commerce applications using a finalized questionnaire while performing set tasks.The procedure was selected as a less expensive way to obtain results about application functionality and design.
- Prediction modeling: Stepwise regression was applied to a sample of 168 observations to model the relationship between ratings and usability predictors.The paper identifies forward selection, backward selection, and stepwise regression as alternative MLR techniques, selecting stepwise for modeling.
- Prediction modeling: K-fold and MRE, including MMRE, were used to assess and validate the proposed rating-prediction model.The model predictions were compared with the hybrid model for rating.
A. SLR ANALYSIS
Simple linear regression evaluated the individual strength of eight usability predictors before multiple linear regression was applied. Efficiency was reported as the most reliable predictor, while human factors were the weakest.
- Regression analysis: Eight usability predictors were first analyzed individually using simple linear regression.The analysis evaluated the individual strength of each possible predictor before multiple linear regression.
- Regression analysis: Efficiency was the most reliable predictor across the usability factors in mobile applications.The result was obtained from the simple linear regression analysis presented in Table 1.
- Regression analysis: Human factors were considered the weakest predictor across the usability factors in mobile applications.The study used regression analysis both for rating prediction and for determining the relative effectiveness of usability factors.
B. MLR ANALYSIS
The study uses multiple linear regression to identify usability factors associated with M-commerce application ratings and selects a seven-factor prediction model. The suggested model explains 91% of rating variation.
- The seven-predictor MLR model achieves R2 = 0.910 for user-rating prediction.The model accounts for 91% variation in determining predicted user ratings.
- The final model includes effectiveness, learnability, communicativeness, consistency, operability, satisfaction, and efficiency.Human factor is excluded from the suggested seven-factor model.
- The regression analysis evaluates combinations of two through eight usability factors as independent variables.Stepwise multilinear regression compares predictor combinations using R2 and significance values.
- The predicted rating is represented by Ŷ in the regression equation.The equation combines the seven selected usability predictors to produce a model-predicted rating.
C. HYBRID VS SUGGESTED MODEL
The study compares the eight-factor hybrid model with a seven-factor regression model and validates the suggested model using fixed and randomly arranged eight-fold evaluations. The reported folds produce significant results with good R2 values.
- The suggested model retains seven of the hybrid model’s eight usability factors, excluding human factor.The hybrid model averages all eight factors for each instance, while the suggested regression model omits HUFA.
- Eight fixed folds divide the data into groups of 21 instances for model evaluation.The fixed groups cover instances 1–21 through 148–168 across K = 1 to K = 8.
- All fixed folds show significant results with good R2 values, and PRED(25) equals one for every fold.The authors report that the fixed-fold evaluation validates the suggested model.
- Randomly arranged eight-fold evaluations also produce significant results with good R2 values.The random-fold analysis is used to assess whether the data remains valid when abruptly rearranged.
VI. CONCLUSION
The conclusion evaluates M-commerce applications through usability factors and develops a factor-based model for rating determination. It reports that several usability factors have a strong impact on ratings and that the stepwise regression identifies a final factor combination.
- The study evaluates M-commerce applications using eight usability factors, each consisting of multiple criteria.The factors are efficiency, learnability, satisfaction, human factor, communicativeness, effectiveness, operability, and consistency.
- Simple linear regression finds significant results with good R2 values for all eight usability factors.The analysis uses the factors as predictors of application ratings.
- Effectiveness, learnability, consistency, and efficiency are reported as having a great impact on rating determination.These factors are identified after applying regression to the eight-factor set.
- Stepwise regression is then used to determine a factor combination for rating prediction.The conclusion begins reporting the selected combination after the stepwise analysis.