Computational models that formalize complex human behaviors enable study and understanding of such behaviors. However, collecting behavior data required to estimate the parameters of such models is often tedious and resource intensive. Thus, estimating dataset size as part of data collection planning (also known as Sample Size Determination) is important to reduce the time and effort of behavior data collection while maintaining an accurate estimate of model parameters. In this article, we present a sample size determination method based on Uncertainty Quantification (UQ) for a specific Inverse Reinforcement Learning (IRL) model of human behavior, in two cases: (1)pre-hocexperiment design—conducted in the planning stage before any data is collected, to guide the estimation of how many samples to collect; and (2)post-hocdataset analysis—performed after data is collected, to decide if the existing dataset has sufficient samples and whether more data is needed. We validate our approach in experiments with a realistic model of behaviors of people with Multiple Sclerosis (MS) and illustrate how to pick a reasonable sample size target. Our work enables model designers to perform a deeper, principled investigation of the effects of dataset size on IRL model parameters.

In this paper, we propose an approach to improve the accuracy of home activity estimation using device-free sensors in the home. This is achieved by transferring existing training data to a new household considering the differences between households. We assumed two scenarios in which we only have training data from other households with labels and we also have training labels for our own household, and proposed the method to compose super-vised transfer between labeled data and unsupervised transfer between labeled unlabeled data for each scenario. To evaluate in realistic settings, we developed the system which consists of an ap-plication for use on tablet terminals, which continuously collect light and any optional sensor data, and a Web-based server system that stores the sensor data, estimates activities, provides visualization on users’ Web browsers, and enables users to edit the activity labels. Using the system, we gathered subjects from open called households during a period of approximately four months, and obtained approximately 11,745 activity inputs, approximately 7.14GB of sensor data, and power consumption data of 237,280 hours from 35 households. As a result of evaluation, our method outperformed naive methods, both in the first and second scenarios.

In this paper, we propose an approach to improve mobile activity recognition, given a training dataset with inaccurate segments, in which the beginning and ending timestamps of homogeneous and continuous activities have inaccurate boundaries due to human errors. In the proposed approach, we A) convert the training dataset to multilabel samples, B) train the dataset by using a multilabel expectation maximization learning algorithm, and C) apply a segmentation method using not only the estimated labels but also the original segment information. We evaluate the proposed approach for three datasets, including simulation data and real activity data, two machine-learning algorithms, and various inaccuracies, and show that the proposed approach outperforms the naive methods as follows: 1) it fixes the segments of the training data and 2) improves the recognition accuracy through cross validation.

In this paper, we provide a real nursing data set for mobile activity recognition that can be used for supervised machine learning, and big data combined the patient medical records and sensors attempted for 2 years, and also propose a method for recognizing activities for a whole day utilizing prior knowledge about the activity segments in a day. Furthermore, we demonstrate data mining by applying our method to the bigger data with additional hospital data. In the proposed method, we 1) convert a set of segment timestamps into a prior probability of the activity segment by exploiting the concept of importance sampling, 2) obtain the likelihood of traditional recognition methods for each local time window within the segment range, and, 3) apply Bayesian estimation by marginalizing the conditional probability of estimating the activities for the segment samples. By evaluating with the dataset, the proposed method outperformed the traditional method without using the prior knowledge by 25.81% at maximum by balanced classification rate. Moreover, the proposed method significantly reduces duration errors of activity segments from 324.2 seconds of the traditional method to 74.6 seconds at maximum. We also demonstrate the data mining by applying our method to bigger data in a hospital.

Non-communicable diseases (NCDs) are no longer just a problem for high-income countries, but they are also a problem that affects developing countries. Preventive medicine is definitely the key to combat NCDs; however, the cost of preventive programs is a critical issue affecting the popularization of these medicine programs in developing countries. In this study, we investigate predictive modeling for providing a low-cost preventive medicine program. In our two-year-long field study in Bangladesh, we collected the health checkup results of 15,075 subjects, the data of 6,607 prescriptions, and the follow-up examination results of 2,109 subjects. We address three prediction problems, namely subject risk prediction, drug recommendation, and future risk prediction, by using machine learning techniques; our multiple-classifier approach successfully reduced the costs of health checkups, a multi-task learning method provided accurate recommendation for specific types of drugs, and an active learning method achieved an efficient assignment of healthcare workers for the follow-up care of subjects.

We present a remote healthcare consultancy system that enables healthcare workers to identify noncommunicable diseases in unreached communities. The healthcare system combines medical sensors with mobile health and is called a Portable Health Clinic. The Portable Health Clinic fits into a briefcase and is operated by the healthcare worker. The goal of this research is to empower the healthcare worker further by allowing her to recognize spurious measurements and to make lifestyle recommendations. In this paper, we show how to process the data: combine, link and compare - captured in patient electronic health records stored in database. We applied association rule technique to find common set of rules in order to build a clinical decision support system. We also showed examples of the meaningful information from the analyzed data to build a better clinical decision support.

Human activity recognition through the wearable sensor will enable a next-generation human-oriented ubiquitous computing. However, most of research on human activity recognition so far is based on small number of subjects, and non-public data. To overcome the situation, we have gathered 4897 accelerometer data with 116 subjects and compose them as HASC2011corpus. In the field of pattern recognition, it is very important to evaluate and to improve the recognition methods by using the same dataset as a common ground. We make the HASC2011corpus into public for the research community to use it as a common ground of the Human Activity Recognition. We also show several facts and results of obtained from the corpus.

In this paper, we introduce a method for recognizing actions in the 1st part of Radio Exercises, with a single smart phone equipped with three-axis accelerometers, which is placed in the breast pocket. Radio Exercises is a warm-up calisthenics performed with music and chants. We obtained the acceleration data of smart phones placed at the breast pockets to a fixed direction of 4 members. For gathered 44 action data, we extracted time windows, and calculated the Fourier transformation for each axis, and adopted the values of from 0.25Hz to 5Hz as a feature vector. Upon the feature vector, we applied principal component extraction, and applied several machine learning algorithms to from 1st to 8th principal components. The benchmarked value resulted in 64.10% of accuracy in the best method, and we could find out what action is easier or difficult to recognize.

In this paper, we propose a secure identification scheme for RFID with efficient time and memory, and also an efficient update of pre-computed values on the server side. Although RFID ( Radio Frequency IDentification ) is becoming popular with pervasive computing environments, a privacy problem still remains, where an adversary can trace users' behavior by linking identification log by adversary readers. For this problem, a hash-chain scheme has been proposed as a secure identification for low-cost RFID tags, and its long identification time has been reduced by Avoine et al. using pre-computation on the server side. However, Avoine's scheme uses static pre-computation, and therefore pre-computed values include ones which are already used and no longer used. An efficient update of pre-computation is needed to remove these wasted values and this can reduce memory usage. In this paper, we optimize a lookup of pre-computed values using d-left hashing, which is a fast lookup technique, and provide efficient update of pre-computed values. We also show reasonable analytical evaluation and experimental result for memory and pre-computation/identification/update time.

As pervasive computing environments become popular, RFID tags are introduced into our daily life. However, there exists a privacy problem that an adversary can trace users' behavior by linking the tag's ID. Although a Hash-chain scheme can solve this privacy problem, the scheme needs a long identification time or a large amount of memory. In this paper, we propose an efficient identification scheme using Bloom filters, which are space-efficient probabilistic data structures. Our Bloom pre-calculation scheme provides a high-speed identification with a small amount of memory by storing pre-calculated outputs of the tags in Bloom filters.

In this paper, we propose an approach to increase the reliability of RFID (radio frequency identification), not relying on the performance of each devices, but relying on a mechanism of entire networked system, which manages objects and/or people in the real world with RFID tags. We identified there exist many factors to reduce the accurate identifications of RFID tags, which results in the reduction of the total-system reliability. Moreover, we show the effect of our approach in diverse parameters such as the number of misidentifications, the length of transportations, the number of RFID readers.