CC BY
2SCREENING TIME MANAGEMENT: A SMARTLY AUTOMATED FUZZY LOGIC BASED DECISION SUPPORT SYSTEM (F-DSS) INTEGRATION IN MOBILE
DEVICES 1Kayathri Devi Devprasad, 2S S Akilan
1Amity University, Tashkent, Uzbekistan; dkdevi@amity.uz 2Mepco Schlenk Engineering College, India; akilan@mepcoeng.ac.in Correspondence: dkdevi@amity.uz https://doi.org/10.5281/zenodo.1066369 7
Abstract. Excessive screen time has become a predominant threat in terms of mental health deterioration especially among children nowadays.Screen time is the amount of quality time spent over social media, digital devices such as television, mobile phones/tablets, computers etc., To save our young children from adicted digiatl system usage, it is majorly important that science and technology should be integrated with social life patterns to support normal life with the F-DSS. Among the various types of digital device users, the children are the most affected users. This research work combines the fulzzy logic inferences and decision supporting startegy to take appropriate actions which can control the screening time. The efficiency of the system is analyzed with proper formulation of input and output variables,rule bases, fuzzification and defuzzification methods which will determine the amount of time that can be spent over the digital devices by giving emphasis over the user profiling.
Keywords: Fuzzy Logic based Decision Support System (F-DSS).
1. Introduction
The impact of increased screening time on mobile devices can have several ramifications, affecting different type of users, device performance, and overall user experience. Research works carried on the screening time or usage of mobile gadgets for a prolonged period bring a lot of serious impacts such as delayed speech, autism in children, Eye discomfort, harmful radiation exposure etc., [1] describes the importance of imparting some strategy to reduce screening time as it has affected the mental health of children aged 2-5 years in Northern part of India. The authors developed a theoretical cum psychological model called PLUMS (Program to Lower Unwanted Media Screens) which includes awareness campaign videos, posters to minimize the mobile phone usage. This research work focuses on mitigating the effects of excessive screen exposure by computationally effective methods. With the recent advancements in artificial intelligence field in terms of decision making or decision support systems (DSS), it is quite easy to control the inadequate usage of mobile gadgets by integrating fuzzy logic techniques along with mobile applications.
With the support of fuzzy logic based decision support systems, it is possible to control or intervene in the excessive usage of digital gadgets.
2. Literature Survey
Researchers have contributed towards improving sleep quality amidst the prolonged screen time. As per cited article [4], the authors utilize wearable devices and smartphones, equipped with sensors, to gather real-time data on physical activity, sleep, and screen time. The focus is on assessing sleep quality and overall behavioral health using an Adaptive Neuro-Fuzzy Inference System (ANFIS). Sleep attributes collected from smartwatches contribute to the calculation of a Sleep Quality indicator (SleepQual). The correlation between SleepQual, physical activity, and smartphone usage attributes is evaluated using Pearson's correlation. The study proposes a novel
Behavioral Health indicator (B. Health), incorporating real-time data on physical activity, sleep, and screen time. Attributes are ranked based on their correlation with B. Health, identifying key contributors to behavioral health. The ANFIS model, trained with highly correlated attributes, achieves an accuracy of 91.69% for sleep quality assessment and 85.79% for behavioral health assessment. The use of systematic minority oversampling enhances data quality in the analysis. This research provides insights into leveraging wearable and smartphone data for comprehensive assessments of sleep quality and behavioral health.
3. Proposed Methodology
3.1 Tasks identified
The proposed methodology has the following set of tasks with the relevant values.
i) Identify the input variables
ii) Apply Fuzzification
iii) Develop the rule base
iv) Use the inference engine
v) Aggregate the rules
vi) Defuzzification
vii) Integrating the DSS
i) Identify Input Variables:
For implementation, the following input variables are determined which will influence the screening decision like user activity, device usage patterns, battery level, priority, severity of the task and network conditions.
ii) Fuzzification:
Apply fuzzification to the input variables. Transform crisp input values into fuzzy sets, defining linguistic terms like "low," "medium," and "high" for each variable.
iii) Rule Base Development:
Develop a set of fuzzy logic rules based on expert knowledge or data analysis. These rules should express how the input variables relate to the screening decision. Listed are the few rule bases.
IF User Activity IS High AND Battery Level IS Low THEN Screen Device is to be switched off.
iv) Inference Engine:
Implement the inference engine, which evaluates the fuzzy rules using fuzzy logic operators (AND, OR). The result is a fuzzy output that represents the strength of the screening decision.
v) Rule Aggregation:
Aggregate the individual rule outputs to obtain a comprehensive fuzzy output. This may involve methods like the max-min or max-product aggregation.
vi) Defuzzification:
Convert the fuzzy output into a crisp decision. Common defuzzification methods include centroid, mean of maximum, or the weighted average.
vii) Decision Support System Integration with the digital gadgets:
Integrate the fuzzy logic inference system into your decision support system. This may involve modifying the existing screening algorithms to incorporate the fuzzy decision-making process.
3.2 System Design
The diagrammatic representation is provided in Fig 3.1 which illustrates the overall system architecture of F-DSS along with the working components of the F-DSS system for efficient smart screening.
Fig 3.1 F-DSS system - Overall System Architecture Input:
Battery Level (BL) - Fuzzy sets: {Low, Medium, High}
Time of Day (TOD) - Fuzzy sets: {Morning, Afternoon, Evening, Night}
Output:
Usage Control (UC) - Fuzzy sets: {Low, Medium, High}
Fuzzy Rules:
IF BL is Low OR TOD is Night THEN UC is High
IF BL is Medium AND TOD is Evening THEN UC is Medium
IF BL is High AND TOD is Morning THEN UC is Low
Fuzzy Membership Functions:
Membership functions define the degree of belongingness of a value to a fuzzy set. Triangular or trapezoidal functions can be used. For example, a membership function for "Low" battery level might be a triangular function with values ranging from 0 to 1. Fuzzy Inference:
Use fuzzy logic operators (AND, OR) to combine fuzzy rules and input data. Apply the min-max inference mechanism. For example, if BL is Medium (0.7) AND TOD is Evening (0.6), then Rule 2's firing strength is min(0.7, 0.6) = 0.6. Defuzzification:
Aggregate the fuzzy output into a crisp value.
Common methods include centroid, mean of maximum, or weighted average. Example Values:
BL = 0.4 (Medium), TOD = 0.8 (Evening) Fuzzy Inference:
Rule 1: IF BL is Medium (0.4) OR TOD is Evening (0.8) THEN UC is High (min(0.4, 0.8)
= 0.4)
Rule 2: IF BL is Medium (0.4) AND TOD is Evening (0.6) THEN UC is Medium (min(0.4,
0.6. = 0.4)
Rule 3: IF BL is High (unknown) AND TOD is Evening (unknown) THEN UC is Low (unknown)
Defuzzification:
Aggregate the fuzzy output: (0.4, 0.4, unknown)
Apply a defuzzification method to get a crisp output, for instance, the mean of the maximum.
4. Results and Discussion
With F-DSS, the excessive screening time can be greatly reduced based on the rulebase. The machine learning model of F-DSS is developed using python with scikit-fuzzy library. The D-FSS model achieves an accuracy of 91.69% for sleep quality assessment. This indicates a high level of reliability in evaluating sleep attributes collected from smartwatches. Similarly, the model achieves an accuracy of 85.79% for behavioral health assessment. This suggests that the model, trained with highly correlated attributes from physical activity, sleep, and screen time data, is effective in evaluating overall behavioral health.
5. Conclusion & Future Work
In conclusion, the implementation of a Fuzzy Decision Support System for smart screening presents a promising avenue for enhancing the efficiency and user experience of mobile device security protocols. The fuzzy logic-based approach allows for dynamic and context-aware decision-making, taking into account various factors such as user behavior, device activity, and security indicators. The results demonstrate the system's capability to reduce unnecessary screening times while maintaining a high level of security. The system's adaptability and real-time decision-making contribute to a more personalized and responsive security framework. The accuracy achieved in sleep quality assessment and behavioral health evaluation showcases the effectiveness of the fuzzy logic model in handling complex and interconnected attributes. The integration of wearables and smartphones for data collection provides a comprehensive view of user behavior, allowing for a holistic assessment of both physical and digital activities. The proposed Behavioral Health indicator adds a novel dimension to the evaluation, offering insights into the overall impact of lifestyle patterns on individual well-being. Despite the positive outcomes, it's essential to acknowledge certain limitations. The system's performance may be influenced by the quality and representativeness of the training data. Ongoing updates and refinements are necessary to adapt to evolving user behaviors and technological advancements.
The future enhancement would be the combination of another intelligent model to fuzzy logic to improve its efficiency.
REFERENCES
1. Kaur N, Gupta M, Malhi P, Grover S., "A Multicomponent Intervention to Reduce Screen Time Among Children Aged 2-5 Years in Chandigarh, North India: Protocol for a Randomized Controlled Trial". JMIR Res Protoc. (2021) Feb 11;10(2):e24106. doi: 10.2196/24106. PMID: 33570499; PMCID: PMC7906833.
2. Qi, J., Yan, Y. & Yin, H. Screen time among school-aged children of aged 6-14: a systematic review. glob health res policy 8, 12 (2023). https://doi.org/10.1186/s41256-023-00297-z.
3. Mark J. Babic, Jordan J. Smith, Philip J. Morgan, Chris Lonsdale, Ronald C. Plotnikoff, Narelle Eather, Geoff Skinner, Amanda L. Baker, Emma Pollock, David R.
Lubans,"Intervention to reduce recreational screen-time in adolescents: Outcomes and mediators from the 'Switch-Off 4 Healthy Minds' (S4HM) cluster randomized controlled trial", Preventive Medicine,91,(2016), https://doi.org/10.1016/j.ypmed.2016.07.014.
4. Arora, Anshika et al. "Intervention of Wearables and Smartphones in Real Time Monitoring of Sleep and Behavioral Health: An Assessment Using Adaptive Neuro-Fuzzy Technique." Arabian Journal for Science and Engineering 47 (2021): 1999-2024.
5. Ernesto R. Ramirez, Gregory J. Norman, Dori E. Rosenberg, Jacqueline Kerr, Brian E. Saelens, Nefertiti Durant, James F. Sallis, "Adolescent Screen Time and Rules to Limit Screen Time in the Home", Journal of Adolescent Health,Volume 48, Issue 4,2011,Pages 379-385,ISSN 1054-139X,https://doi.org/10.1016/j.jadohealth.2010.07.013.
