Mobile Device Environment Research Articles

Modified bi-directional long short-term memory and hyperparameter tuning of supervised machine learning models for cardiovascular heart disease prediction in mobile cloud environment

In recent decades, cardiovascular heart disease (CHD) has emerged as the predominant cause of mortality on a global scale. Numerous risk factors for cardiovascular disease necessitate prompt access to accurate, dependable, and efficient early diagnosis and disease management strategies. The timely and precise identification of cardiac disease holds significant importance in the prevention and treatment of heart failure. The reliability of diagnosing heart disease by standard medical history has been widely questioned. Medical imaging, including angiography, is the most widely used method for diagnosing CAD. Nevertheless, angiography is renowned for its high cost and its correlation with many adverse consequences. Noninvasive techniques, such as machine learning, demonstrate reliability and efficiency in the classification of individuals as either healthy or afflicted with cardiac disease. In this research, a new efficient real-time diagnostic method for CHD was designed via hyperparameter tuning of supervised machine learning (SML). This study consists of five significant components: the data preprocessing stage, classification with and without hyperparameter tuning, model integration and clinical decision, intervention and patient engagement, and lastly, model deployment and API integration on Amazon Cloud SageMaker. The data was preprocessed, which included finding outliers using the interquartile range (IQR), normalizing the data, and using the synthetic minority oversampling method (SMOTE) to fix the imbalance in the data. Following this, we propose a new adaptation of the Bi-Directional Long-Short Term Memory (BiLSTM) model, incorporating an attention mechanism (AM) to facilitate feature selection. The feature relevance was determined by calculating the coefficient scores for the top ten (10) features in each of the SMLs, namely the Catboost, extreme gradient boosting (XgBoost), random forest (RF), and linear discriminant analysis (LDA). Both with and without the inclusion of hyperparameter tuning, the grid search process further improved the models. We deployed our proposed models at scale by outsourcing them to the SageMaker cloud environment. The detailed experimental analysis showed that the BiLSTM + AM-XgBoost model worked best when tuned for hyperparameters via the grid search method. Specifically, this model achieved an accuracy of 99.99%, recall of 99.65%, precision of 99.68%, and AUC of 88.16%. The BiLSTM + AM-XgBoost model, without any hyperparameter adjustment, achieved an accuracy of 97.58%, recall of 96.32%, precision of 97.32%, and AUC of 87.32%. The tuning of hyperparameters affects the performance of SML in terms of accuracy, recall, precision, and AUC. In addition, we conducted a comparative analysis of our proposed models against the current state-of-the-art approaches. The findings of our evaluation indicate that our models exhibit a level of performance that outperforms the existing methods. The proposed model can function as a decision support system that employs SML techniques and seeks to enhance the efficacy of cardiac patient diagnostics for healthcare professionals. The simplicity of the suggested model is attributed to its reduced computational complexity. The findings demonstrate the potential application of our method in the field of telemedicine, particularly in the context of mobile cloud device environments for CHD.

РАЗРАБОТКА МОБИЛЬНОГО ПРИЛОЖЕНИЯ НА БАЗЕ ОС ANDROID ДЛЯ ВЗАИМОДЕЙСТВИЯ СТУДЕНТОВ И ПРЕПОДАВАТЕЛЕЙ

In the course of the work, a mobile messenger application is being developed for the possibility of remote interaction between students and professors. Application development is carried out in the Android Studio development environment for mobile devices running the Android operating system

A Novel Hybrid Deep Learning Approach to Code Generation Aimed at Mitigating the Real-Time Network Attack in the Mobile Experiment Via GRU-LM and Word2vec

As the use of devices in mobile environments increases, network attacks such as DDoS have a malicious attempt to flood the network's regular traffic to overload the target and surrounding infrastructure. This research proposed machine learning and deep learning approaches to dealing with DDoS attacks, and the results are described as follows. First, this research successfully detected DDoS attacks through an LGBM with a 100% accuracy score. Second, the proposed model (GRU-LM), which consists of a trained Word2vec layer with the Python dataset, is far more effective than the standard GRU model. Since Python is quite similar to English, language model-based GRU yields superior results. Various preprocessing steps were performed through the NLTK package, and each number was assigned to the tokenized one for constructing the GRU language model. The result reveals that the proposed model achieved an accuracy score of 87% for predicting the following words in the source code, while the rest achieved below 30% accuracy. This conclusion is significant because its relatively simple and light structure overcomes tradeoff problems between time and accuracy and is adaptable to the mobile setting. Discovering traffic patterns for the underlying data of DDOS assaults and retrieving them using statistical data analysis is the value of this research. Furthermore, since public cloud application vulnerability assaults are rising due to expanding cloud infrastructure, this finding could be used in such attacks.

Recurrent Neural Network-Based Multimodal Deep Learning for Estimating Missing Values in Healthcare

This estimation method operates by integrating the input values that are redundantly collected from heterogeneous devices through the selection of a representative value and estimating missing values by using a multimodal RNN. Users use a heterogeneous healthcare platform mainly in a mobile environment. Users who pay a relatively large amount of attention to healthcare possess various types of healthcare devices and collect data through their mobile devices. The collected data may be duplicated depending on the types of these devices. This data duplication causes an ambiguity issue in that it is difficult to determine which value among multiple data should be taken as the user’s actual value. Accordingly, it is necessary to create a neural network structure that considers the data value at the time previous to the current time. RNNs are appropriate for handling data with a time series characteristic. To learn an RNN-based neural network, learning data that have the same time step are required. Therefore, an RNN in which one variable becomes single-modal was designed for each learning run. In the RNN, a cell is a gated recurrent unit (GRU) cell that presents sufficient accuracy in the small resource environment of mobile devices. The RNNs that are learned according to the variables can each operate without additional learning, even if the situation of the user’s mobile device changes. In a heterogeneous environment, missing values are generated by various types of errors, including errors caused by battery charge and discharge, sensor failure, equipment exchange, and near-field communication errors. The higher the missing value ratio, the greater the number of errors that are likely to occur. For this reason, to achieve a more stable heterogeneous health platform, missing values must be considered. In this study, a missing value was estimated by means of multimodal deep learning; that is, a multimodal deep learning method was designed with one neural network that was connected with each learned single-modal RNN using a fully connected network (FCN). Each RNN input value delivers mutual influence through the weights of the FCN, and thereby, it is possible to estimate an output value even if any one of the input values is missing. According to the evaluation in terms of representative value selection, when a representative value was selected by using the mean or median, the most stable service was achieved. As a result of the evaluation according to the estimation method, the accuracy of the RNN-based multimodal deep learning method is 3.91%p higher than that of the SVD method.

Hurdles for adopting mobile learning devices at the outset of clinical courses

BackgroundMobile devices provide medical students with easy access to medical information and educational resources. Since 2013, we have followed the study use of iPads among medical students. In 2016, we observed a notable drop in the mobile device usage in the first cohort of medical students entering their clinical courses.MethodsThe aim of the study was to identify the hurdles for adopting mobile devices at the beginning of the clinical courses. We examined how students evaluated their own and the clinical teachers’ ability to use the iPad, how the study assignments fit into digital learning, and how students used the mobile device with patients. The data were collected with online surveys among three consecutive student cohorts and the distributions of closed-ended questions analyzed.ResultsResponse rates ranged from 67.5 to 90.8%. Students evaluated their own ability to use the iPad as good or excellent and teachers’ skills as relatively poor and wanted more digitally tailored assignments. They reported negative attitudes towards mobile device use in the clinical setting and were hesitant to use them in patient contact. Teachers seldom communicated suitable quality medical applications to students.ConclusionsClinical teachers need support and training to implement a learning environment and assignments appropriate for mobile devices. Both students and teachers were concerned about using these devices with patients. To achieve the full potential of digitalisation in clinical courses, their use should be developed collectively with students.

Quantum Game Application to Recovery Problem in Mobile Database

Mobile Computing (MC) is a relatively new concept in the world of distributed computing that is rapidly gaining traction. Due to the dynamic nature of mobility and the limited bandwidth available on wireless networks, this new computing environment for mobile devices presents significant challenges in terms of fault-tolerant system development. As a consequence, traditional fault-tolerance techniques are inherently inapplicable to these systems. External circumstances often expose mobile systems to failures in communication or data storage. In this article, a quantum game theory-based recovery model is proposed in the case of a mobile host’s failure. Several of the state-of-the-art recovery protocols are selected and analyzed in order to identify the most important variables influencing the recovery mechanism, such as the number of processes, the time needed to send messages, and the number of messages logged-in time. Quantum game theory is then adapted to select the optimal recovery method for the given environment variables using the proposed utility matrix of three players. Game theory is the study of mathematical models of situations in which intelligent rational decision-makers face conflicting interests (alternative recovery procedures). The purpose of this study is to present an adaptive algorithm based on quantum game theory for selecting the most efficient context-aware computing recovery procedure. The transition from a classical to a quantum domain is accomplished in the proposed model by treating strategies as a Hilbert space rather than a discrete set and then allowing for the existence of linear superpositions between classical strategies; this naturally increases the number of possible strategic choices available to each player from a numerable to a continuous set. Numerical data are provided to demonstrate feasibility.

Corrections to “HMO: Ordering RFID Tags With Static Devices in Mobile Environments”

Presents corrections to the acknowledgement section for the above named article.

Mobile Cloud Computing as Mobile offloading Solution: Frameworks, Focus and Implementation Challenges

Mobile devices have been operating under limited resource capacity including memory, processing speed, storage and battery life. Due to the advancement of technology, complex applications have been designed and implemented, and these application demand computing device with high capacity. Cloud computing emerged as solution for computing device with limited capacity. However, integrating mobile operating environment with cloud computing has been a challenge due to dynamicity of mobile device environment including unreliability of wireless communication. This paper reviews recent studies in Mobile Cloud Computing to assess its implementation as one of task offloading solution for mobile device. The study reviewed common framework that are used to implement Mobile Cloud Computing, the focus of the current studies and finally highlighted open issues that need to be addressed when implementing optimal Mobile Cloud Computing. Keywords: Mobile Cloud Computing, task offloading, Mobile device, cloudlet DOI: 10.7176/CEIS/11-5-04 Publication date:September 30th 2020

Mobile Cloud Forensic Readiness Process Model for Cloud-Based Mobile Applications

Nowadays, mobile cloud applications have attracted millions of smartphone users due to the proliferation of cyber technologies and a wide range of applications. Mobile cloud forensic investigation methodologies need tremendous growth due to the increasing crime incidents. The forensic readiness model plays a significant role in the forensic investigation framework by ensuring the ease of investigation to the forensic investigator. The existing forensic readiness supports either the mobile device or cloud environment, which lacks to prepare the information for the investigation. This article presents a mobile cloud forensic readiness process model to identify the factors and prepare the information that effectively supports forensic investigations. The proposed model involves requirements for the mobile cloud forensics from multiple perspectives with the aim of developing the forensic-ready system. As a result, the proposed forensic readiness model enables the mobile cloud forensic to improve the accuracy of investigation as well as reduces the investigation time significantly.

SBADR: stable and bandwidth aware dynamic routing protocol for mobile ad hoc network

Purpose Quality communication is a big challenge in mobile ad hoc networks because of a restricted environment for mobile devices, bandwidth-constrained radio connections, random mobility of connected devices, etc. High-quality communication through wireless links mainly depends on available bandwidth, link stability, energy of nodes, etc. Many researchers proposed stability and link quality methods to improve these issues, but they still require optimization. This study aims to contribute towards better quality communication in temporarily formed networks. The authors propose the stable and bandwidth aware dynamic routing (SBADR) protocol with the aim to provide an efficient, stable path with sufficient bandwidth and enough energy hold nodes for all types of quality of service (QoS) data communication. Design/methodology/approach The proposal made in this work used received signal strength from the media access control (MAC) layer to estimate the stability of the radio connection. The proposed path stability model combines the stability of the individual link to compute path stability. The amount of bandwidth available for communication at a specific time on a link is defined as the available link bandwidth that is understood as the maximum throughput of that link. Bandwidth as a QoS parameter ensures high-quality communication for every application in such a network. One other improvement, towards quality data transmission, is made by incorporating residual energies of communicating and receiving nodes in the calculation of available link bandwidth. Findings Communication quality in mobile ad hoc network (MANET) does not depend on a single parameter such as bandwidth, energy, path stability, etc. To address and enhance quality communication, this paper focused on high impact factors, such as path stability, available link bandwidth and energy of nodes. The performance of SBADR is evaluated on the network simulator and compared with that of other routing protocols, i.e. route stability based QoS routing (RSQR), route stability based ad-hoc on-demand distance vector (RSAODV) and Ad-hoc on-demand distance vector (AODV). Experimental outcomes show that SBADR significantly enhanced network performance in terms of throughput, packet delivery ratio (PDR) and normalized control overhead (NCO). Performance shows that SBADR is suitable for any application of MANET having random and high mobility. Research limitations/implications QoS in MANET is a challenging task. To achieve high-quality communication, the authors worked on multiple network parameters, i.e. path stability, available link bandwidth and energy of mobile nodes. The performance of the proposed routing protocol named SBADR is evaluated by a network simulator and compared with that of other routing protocols. Statistical analysis done on results proves significant enhancement in network performance. SBADR is suitable for applications of MANET having random and high mobility. It is also efficient for applications having a requirement of high throughput. Practical implications SBADR shows a significant enhancement in received data bytes, which are 1,709, 788 and 326 more in comparison of AODV, RSAODV and RSQR, respectively. PDR increased by 21.27%, 12.1%, 4.15%, and NCO decreased by 9.67%, 5.93%, 2.8% in comparison of AODV, RSAODV and RSQR, respectively. Social implications Outcomes show SBADR will perform better with applications of MANET such as disaster recovery, city tours, university or hospital networks, etc. SBADR is suitable for every application of MANET having random and high mobility. Originality/value This is to certify that the reported work in the paper entitled “SBADR: stable and bandwidth aware dynamic routing protocol for mobile ad hoc network” is an original one and has not been submitted for publication elsewhere. The authors further certify that proper citations to the previously reported work have been given and no data/tables/figures have been quoted verbatim from the other publications without giving due acknowledgment and without permission of the author(s).

HMO: Ordering RFID Tags with Static Devices in Mobile Environments

Passive Radio Frequency Identification (RFID) tags have been widely applied in many applications, such as logistics, retailing, and warehousing. In many situations, the order of objects is more important than their absolute locations. However, state-of-art ordering methods need a continuing movement of tags and readers, which limit the application domain and scalability. In this paper, we propose a 2-dimension ordering approach for passive tags that requires no device movement. Instead, our method utilizes signal changes caused by arbitrary movement of human beings around tags, who carry no device for horizontal dimension ordering. Hence, our method is called Human Movement based Ordering (HMO). The basic idea of HMO is that when people pass between the reader antenna and tags, the received signal strength will change. By observing the time-series RSS changes of tags, HMO can obtain the order of tags along with a specific horizontal direction. For vertical dimension, we employ a linear programming method that is tolerant of tiny errors in practice. We implement HMO with commodity off-the-shelf RFID devices. The experimental results show that HMO can achieve up to 88.71 and 90.86 percent average accuracies in the signal- and multi-person cases, respectively.

Technology for providing electromagnetic compatibility requirements for mobile wireless systems

The subject of study is the process of ensuring electromagnetic compatibility requirementsof mobile wireless communication systems. The goal is the development of ultra-wideband communication technology, which provides a reduction in the level of interference electromagnetic environment of mobile devices, occupying a random position relative to each other. The technology is based on the principle of direct transmission of low-power encoded pulses in a very wide frequency band without a carrier frequency. The task is to reduce the level of interference electromagnetic environment and compensate for distortions of the information signal caused by intersymbol interference and multipath radio wave propagation. Methods used: methods of analytical, simulation and temporary positional-pulse coding. The following results are obtained. A technology has been developed to ensure the electromagnetic compatibility requirements of mobile wireless communication systems, which includes the method of temporary position-pulse coding and the method of shimmering polarization of information signals. The minimum required number of pulses encoding an information bit is proved, which ensure its recovery. A technical solution is proposed for the design of the antenna system for signals with shimmering polarization. The possibility of joint noiseless operation in the same frequency range of both traditional communication systems and systems using ultra-wideband signals is shown. Conclusions. The use of the proposed technology of ultra-wideband signals, shimmering polarization of signals and the antenna system with the communication systems of mobile devices allows increasing the signal-to-noise ratio at the receiver input, which reduces the level of electromagnetic radiation, thus providing the electromagnetic compatibility requirements of mobile wireless communication systems. Efficiency of the developed technology application is conditioned by the possibility of realizing a high density of communication channels per square meter of the working area. By reducing the level of the information signal relative to white noise in the operating frequency range, simultaneous noise-free operation of both narrow-band traditional communication systems and communication systems using ultra-wideband signals is carried out. Implementation of the technology for ensuring the electromagnetic compatibility requirements of mobile wireless communication systems makes it possible to ensure their implementation at all stages of the development and implementation of mobile wireless communication systems.

Secure Service Composition with Quantitative Information Flow Evaluation in Mobile Computing Environments.

The advances in mobile technologies enable mobile devices to cooperate with each other to perform complex tasks to satisfy users’ composite service requirements. However, data with different sensitivities and heterogeneous systems with diverse security policies pose a great challenge on information flow security during the service composition across multiple mobile devices. The qualitative information flow control mechanism based on non-interference provides a solid security assurance on the propagation of customer’s private data across multiple service participants. However, strict discipline limits the service availability and may cause a high failure rate on service composition. Therefore, we propose a distributed quantitative information flow evaluation approach for service composition across multiple devices in mobile environments. The quantitative approach provides us a more precise way to evaluate the leakage and supports the customized disciplines on information flow security for the diverse requirements of different customers. Considering the limited energy feature on mobile devices, we use a distributed evaluation approach to provide a better balance on consumption on each service participant. Through the experiments and evaluations, the results indicate that our approach can improve the availability of composite service effectively while the security can be ensured.

A New Deep Learning-Based Handwritten Character Recognition System on Mobile Computing Devices

Deep learning (DL) is a hot topic in current pattern recognition and machine learning. DL has unprecedented potential to solve many complex machine learning problems and is clearly attractive in the framework of mobile devices. The availability of powerful pattern recognition tools creates tremendous opportunities for next-generation smart applications. A convolutional neural network (CNN) enables data-driven learning and extraction of highly representative, hierarchical image features from appropriate training data. However, for some data sets, the CNN classification method needs adjustments in its structure and parameters. Mobile computing has certain requirements for running time and network weight of the neural network. In this paper, we first design an image processing module for a mobile device based on the characteristics of a CNN. Then, we describe how to use the mobile to collect data, process the data, and construct the data set. Finally, considering the computing environment and data characteristics of mobile devices, we propose a lightweight network structure for optical character recognition (OCR) on specific data sets. The proposed method using a CNN has been validated by comparison with the results of existing methods, used for optical character recognition.

How information processing and mobile channel choice influence product returns: An empirical analysis

AbstractThe mobile internet is starting to overtake the desktop device‐based internet as a purchase channel. Its impact on consumer behavior is therefore increasingly important to understand. This study seeks to understand and measure, if usage of mobile devices for online purchases leads to a lower decision quality and, in effect, to more product returns. In doing so, the impact of information environments on the end‐to‐end consumer purchase decision‐making process is better understood and it is investigated, if the information environment of mobile devices leads consumers to take more error‐prone purchase decisions. An exclusive data set spanning more than 140 million transactions of a European online retailer is used to empirically analyze changes in product return behavior after mobile channel adoption. The results show that mobile channel usage is positively related to product returns, overall and for both, purchases made with mobile devices and purchases made with desktop devices, although prior literature predicts that returns from desktop purchases should not increase. These findings suggest that through new channels, consumers’ information environment is altered sufficiently to affect their decision accuracy. Moreover, the results indicate that previous research may be overestimating the positive effect of mobile channel adoption on sales by disregarding changes in product return behavior.

PrivateZone: Providing a Private Execution Environment Using ARM TrustZone

ARM TrustZone is widely used to provide a Trusted Execution Environment (TEE) for mobile devices. However, the use of TrustZone is limited because TrustZone resources are only available for some pre-authorized applications. In other words, only alliances of the TrustZone OS vendors and device manufacturers can use TrustZone to secure their services. To help overcome this problem, we designed the PrivateZone framework to enable individual developers to utilize TrustZone resources. Using PrivateZone, developers can run Security Critical Logics (SCL) in a Private Execution Environment (PrEE). The advantage of PrivateZone is its leveraging of TrustZone resources without undermining the security of existing services in the TEE. To guarantee this, PrivateZone creates a PrEE using a memory region that is isolated from both the Rich Execution Environment (REE) and TEE. In this paper, we describe the design and implementation of PrivateZone. The prototype of PrivateZone was implemented on an Arndale board with a Cortex-A15 dual-core processor. We built PrivateZone by exploring both security and virtualization extensions of the ARM architecture. To illustrate the usage and the efficacy of PrivateZone, we developed an Android application based on PrivateZone framework, and evaluated the performance overhead imposed on the OS in the REE and SCLs in the PrEE.

Blockchain-based mobile fingerprint verification and automatic log-in platform for future computing

Door locks and user authentication are the major issues in the current banking industry. There are a number of user future computing techniques, and security is especially essential to these methods. Existing digital door locks have the problem of opening easily with a stolen PIN number or by electrical shock. Thus, this study proposed and implemented an integrated automatic log-in platform based on mobile fingerprint recognition by applying the blockchain theory. As a result of this research, a convenient, integrated automatic log-in platform with powerful security has been constructed using a smartphone-based fingerprint recognition function. There are three major functions of the platform. First, it is possible to authenticate the user in PC, mobile device, and IoT environments through fingerprint recognition. Second, the platform includes SDK to develop application software for user authentication and IoT services. Last is its strengthened security using the blockchain theory to prepare against tampering/forging/leaking of a user’s fingerprint information by hackers.

Partially shaded sketch-based image search in real mobile device environments via sketch-oriented compact neural codes

With the advent of touch screens in mobile devices, sketch-based image search is becoming the most intuitive method to query multimedia contents. Traditionally, sketch-based queries were formulated with hand-drawn shapes without any shades or colors. The absence of such critical information from sketches increased the ambiguity between natural images and their sketches. Although it was previously considered too cumbersome for users to add colors to hand-drawn sketches in image retrieval systems, the modern day touch input devices make it convenient to add shades or colors to query sketches. In this work, we propose deep neural codes extracted from partially colored sketches by an efficient convolutional neural network (CNN) fine-tuned on sketch-oriented augmented dataset. The training dataset is constructed with hand-drawn sketches, natural color images, de-colorized, and de-texturized images, coarse and fine edge maps, and flipped and rotated images. Fine-tuning CNN with augmented dataset enabled it to capture features effectively for representing partially colored sketches. We also studied the effects of shading and partial coloring on retrieval performance and show that the proposed method provides superior performance in sketch-based large-scale image retrieval on mobile devices as compared to other state-of-the-art methods.

Unobtrusive touch‐free interaction on mobile devices in dirty working environments

AbstractIn this work, we propose a new mode of interaction using hand gestures captured by the back camera of a mobile device. Using a simple intuitive two‐finger picking gesture, the user can perform mouse click and drag‐and‐drop operations from the back of the device, which provides an unobstructed touch‐free interaction. This method allows users to operate on a tablet and maintain full awareness of the display, which is especially suitable for certain working environments, such as the machine shop, garage, kitchen, gym, or construction site, where people's hands may be dirty, wet, or in gloves. The speed, accuracy, and error rate of this interaction have been evaluated and compared with the typical touch interaction in an empirical study. The results show that, although this method is not, in general, as efficient and accurate as direct touch, participants still consider it an effective and intuitive method of interacting with mobile devices in environments where direct touch is impractical.

A heuristics approach to mine behavioural data logs in mobile malware detection system

Nowadays, in the era of Internet of Things when everything is connected via the Internet, the number of mobile devices has risen exponentially up to billions around the world. In line with this increase, the volume of data generated is enormous and has attracted malefactors who do ill deeds to others. For hackers, one of the popular threads to mobile devices is to spread malware. These actions are very difficult to prevent because the application installation and configuration rights are set by owners, who usually have very low knowledge or do not care about the security. In this study, our aim is to improve security in the environment of mobile devices by proposing a novel system to detect malware intrusions automatically. Our solution is based on modelling user behaviours and applying the heuristic analysis approach to mobile logs generated during the device operation process. Although behaviours of individual users have a significant impact on the social cyber-security, to achieve the user awareness has still remained one of the major challenges today. For this task, there is proposed a light-weight semantic formalization in the form of physical and logical taxonomy for classifying the collected raw log data. Then a set of techniques is used, like sliding windows, lemmatization, feature selection, term weighting, and so on, to process data. Meanwhile, malware detection tasks are performed based on incremental machine learning mechanisms, because of the potential complexity of this tasks. The solution is developed in the manner to allow the scalability with several blocks that cover pre-processing raw collected logs from mobile devices, automatically creating datasets for machine learning methods, using the best selected model for detecting suspicious activity surrounding malware intrusions, and supporting decision making using a predictive risk factor. We experimented cautiously with the proposal and achieved test results confirm the effectiveness and feasibility of the proposed system in applying to the large-scale mobile environment.