Skip to main content
. 2022 May 25;9(2):e29767. doi: 10.2196/29767

Table 2.

Characteristics of phase 1 studies (N=52).

Characteristics Phase 1 studies, n (%)
Number of surveyed studies or apps

Android 45 (87)

iOS 6 (12)

Both 1 (2)
Roles of MLa

Recognition 19 (37)

Detection 6 (12)

Prediction 4 (8)

Recognition and monitoring 4 (8)

Recognition and recommendation 4 (8)

Recognition and estimation 2 (4)

Recommendation and monitoring 1 (2)

Recommendation 1 (2)

Estimation 1 (2)

Recognition, recommendation, and monitoring 1 (2)
Types of ML

Supervised learning 39 (75)

Unsupervised learning 1 (2)

Both 1 (2)

External ML library 2 (4)
Processing techniques

Data 25 (48)

Image 11 (21)

Image and calculation 4 (8)

Data and calculation 4 (8)

Voice 3 (6)

Calculation 3 (6)

Image, data, and calculation 2 (4)
Focus

Physical health 19 (37)

Weight control 14 (27)

Disease 9 (17)

Mental health 5 (10)

Sleep 3 (6)

Recipe’s recommendation 1 (2)

Multidimensional 1 (2)
Crucial functionalities

Recognition 20 (38)

Detection 9 (17)

Prediction 4 (8)

Recognition and recommendation 4 (8)

Recognition and monitoring 4 (8)

Recommendation and monitoring 3 (6)

Monitoring 3 (6)

Recognition and estimation 2 (4)

Estimation 1 (2)

Recommendation 1 (2)

Recognition, recommendation, and monitoring 1 (2)
UIb components

Label 18 (35)

Image 17 (33)

Button 15 (29)

Input box 8 (15)

List 8 (15)
Navigation and structure

Tab 6 (12)

Back and next 5 (10)

Main page 2 (4)

Tab and back and next 1 (2)

Tab, back and next, and hamburger menu 1 (2)
Services and technologies

Motion sensors 21 (40)

Camera 18 (35)

GPS 2 (4)

Microphone 4 (8)
Security features

Log-in 1 (2)
Architectures and patterns

Client-server (web-based) 30 (58)

On device (offline) 19 (37)

MVCc 2 (4)
Inference of ML

Web-based inference 18 (35)

Offline inference 10 (19)

Both 3 (7)

Web-ready solutions 2 (4)
Development approach

Prototype 8 (15)

User-centered design 2 (4)

Agile 1 (2)

Extreme programming 1 (2)

Iterative 1 (2)
Logging mechanisms

Automatic 24 (46)

Manual 24 (46)

Both 4 (8)
Evaluation

Algorithm’s performance 30 (58)

Algorithm’s accuracy 9 (17)

Algorithm’s performance and cross-validation 8 (15)

Usability study 4 (8)

Cross-validation 1 (2)

aML: machine learning.

bUI: user interface.

cMVC: model-view-controller.