新居 学

Total knee arthroplasty (TKA) is one of the common knee surgeries. Because there are some types of TKA implant, it is hard to select appropriate type of TKA implant for individual patient. For the sake of pre-operative planning, this study presents a novel approach, which predicts post-operative implanted knee function of individuals. It is based on a clinical big data analysis. The big data is composed by a set of pre-operative knee mobility function and post-operative knee function. The method constructs a post-operative knee function prediction model by means of a machine learning approach. It extracts features using principal component analysis, and constructs a mapping function from pre-operative feature space to post-operative feature space. The method was validated by applying to prediction of post-operative anterior-posterior translation in 52 TKA operated knees. Leave-one-out cross validation test revealed the prediction performances with a mean correlation coefficients of 0.79 and a mean root-mean-squared-error of 3.44 mm.

In recent years, medical institutions have very big data including medical images. The big image data analysis using the collected medical images is effective to increase the accuracy and the reproducibility of the surgery. Anterior cruciate ligament (ACL) injury causes knee joint instability, and affects on sports performance. Therefore, ACL reconstruction surgery is essential to keep their performance high and to prevent osteoarthrosis. We have proposed a MR image based pre-operative planning system of ACL reconstruction. The system manually applies the Quadrant method to the synthesized pseudo radiograph. This paper proposes a fully automated pre-operative planning system based on the clinical big image data analysis. The experimental results showed that the proposed method successfully estimated the bone tunnel opening site to insert the ACL.

Daily healthcare is very important for our quality of life. Especially, in the aging countries like Japan, medical costs will increase in the near future. Almost all Japanese people have health examination in every year. By the health examination, Japanese people understand own health condition. If the results of health examination get worse, then we will have precise examinations to find a cause of the deterioration of the health condition. On the other hand, several health monitoring devices are available. Most of such devices are provided for assisting exercise. We have developed a MEMS-based human condition monitoring device and fuzzified neural network based condition estimation system. The monitoring device can obtain some kinds of sensor data from a subject person. The fuzzified neural network based condition estimation system can estimate the heart rate of the subject person from the obtained sensor data. In this paper, we improve the fuzzified neural network based system for obtaining appropriate estimation results.

In this paper, an application of self-organizing maps (SOM's) in classifying and predicting data of female subjects with unhealthy-level visceral fat is discussed. The proposed method chooses subjects fulfilling the standard specified by body mass index and abdominal circumference. It defines the class with subjects of which hemoglobin A1c (HbA1c) values and item values associated with a liver deteriorate, that with subjects having HbA1c and triglyceride values deteriorate, and that with remaining subjects. Normal SOM learning is conducted, using data generated from original values of twelve items such as HbA1c and glutamic-oxaloacetic. The constructed map consists of neurons with labels. The label of a winner determines the class of the presented unknown data. The prediction depends on the label of a winner for the presented unknown data, a set of original data that determine the label, and a set of next year's data of the subjects with the above original data. Experimental results reveal that the proposed method achieves the reasonably favorable accuracies in classifying data and in predicting HbA1c values.

It is known that lifestyle habit and genetic factor are main reasons that occur cerebral aneurysms. In addition, some studies suggest that cerebral artery shape might be correlated with a risk of occurring aneurysms. For the purpose of preemptive medical care of the cerebral aneurysm, this study proposes a method to estimate a risk of occurring cerebral aneurysms based on the cerebral artery structure. The method extracts morphometric features of the Wills ring such as 3-D artery shape and bifurcation angle in 3-D magnetic resonance angiography (MRA) images. It then estimates the risk of occurring cerebral aneurysms from the extracted features using support vector machines (SVM). To validate the proposed method, we employed 40 subjects with cerebral aneurysms, and 40 subjects without cerebral aneurysms. Leave-one-out cross validation test was performed, and the method using 3-D artery shape achieved a sensitivity of 75% and a specificity of 75%; one using bifurcation angle did a sensitivity of 33% and a specificity of 71%; one using all features did a sensitivity of 68% and a specificity of 89%. The results showed that 3-D shape is effective for cerebral aneurysm occurrence risk prediction.

In the infertility treatment, medical examinations using ultrasonic devices, which can diagnose the mother's body safely on real time, are major. It is very difficult to judge the existence of an ovum before carrying out the paracentesis. A system which can distinguish the existence of the ovum in the ovarian follicle using ultrasonic devices is required. In this paper, several features of the deformation of ovarian follicle are defined and extracted from the ultrasonic moving image in a paracentesis operation. These features are extracted from the ultrasonic moving image obtained at the time of an ovum extraction operation. We investigate whether some clusters according to the existence of the ovum are formed using the defined features. Moreover, we also investigate whether some tendency exists in these features by the existence of an ovum.

Osteoarthrosis is a kind of knee disease, and it decreases the patients' Quality of Life. The total knee arthroplasty surgery replaces the damaged knee joint with artificial knee joint. The 3-D kinematics of implanted knee is effective to diagnose and improve the function of the artificial knee joint. The still image analysis based conventional knee kinematics analysis methods do not provide the smooth knee kinematics. This paper proposes the knee kinematics analysis method using particle filter. The experimental results showed that the proposed method successfully estimated the 3-D implanted knee kinematics, and the distance transformation based new likelihood improved the applicability.

Total knee arthroscopy (TKA) is a very effective surgery for damaged knee joint treatment. Because, there are some TKA operation methods and TKA implant products, it is difficult to decide an appropriate one at the pre-operative planning. This study introduces a novel approach to assist surgeon for the pre-operative planning, and proposes a prediction method of post-operative knee joint kinematics. The method is based on principal component analysis (PCA) for characteristics extraction, and machine learning algorithms. The proposed method was validated by leave-one-out cross validation test in 46 osteoarthritis (OA) knee patients. The results show that the proposed method can predict the post-operative knee joint kinematics from the pre-operative one with a mean correlation coefficient of 0.69, and a root-mean-squared-error (RMSE) of 1.8 mm.

We propose rarity-oriented retrieval methods for serendipity using two approaches. We define rare information as relevant and atypical information. We propose two approaches. In the first approach, we use social bookmark data. We introduce tag estimation to our previous work. The second approach is based on word co-occurrence in a dataset. In both approaches, we use conditional probabilities to express relevancy and atypicality. In experiments, we compared our methods with the relevance-oriented method, the diversity-oriented method, and another rarity-oriented method. Our methods using word co-occurrence obtained better nDCG scores than the other methods.

In this paper, we discuss a classification method of nursing-care texts using the word2vec [1]. The word2vec is a tool which provides the continuous bag-of-words and skip-gram implementations for realizing word vectors. We have tackled to classify nursing-care texts, which are freestyle Japanese texts, for improving nursing quality in several years. Several machine learning methods have been used for classifying such texts. To train a machine learning method, we used a word list which contains words appeared in the training data. Since the word list is a mere list, the relation among words is not considered. Also the length of the list depends on the number of words. Word vector representation proposed in [2]-[4] realized word representations in arbitrary dimensional space. We use the word2vec as a alternative word list in this paper. And we propose a new feature vector definition which is based on dependency structures in a text. From experimental results, we compare the proposed definition with our previous works.

In this paper, a method of classifying the data of female subjects taking the specific health examination is presented, using self-organizing maps (SOM's). The proposed method focuses on female subjects fulfilling the standard specified by body mass index and abdominal circumference. It defines the following data classes: the class with subjects of which hemoglobin Alc (HbAlc) values and item values associated with a liver belong to unhealthy levels, that with subjects having HbAlc and triglyceride values included in unhealthy levels, and that with remaining subjects. It generates the data for normal SOM learning from values of twelve items such as HbAlc and glutamic-oxaloacetic. SOM learning is made to construct the map, and its neurons are labeled. The class of data to be checked depends on the label of a winner, when the data is presented. Experimental results establish that the proposed method achieves the reasonably favorable accuracy of consistency on data classification.

Due to inherent characteristic of having priori-information about the shape and appearance, statistical shape models (SSMs) are considered as powerful tools in 3D-MR image analysis. Such as, the SSMs of femoral bone can be used for quantification in knee surgeries, in particular, automated segmentation of femoral bony region to be applied in computer-aided surgical planning of anterior cruciate ligament (ACL) reconstruction. This paper mainly focuses on a method of automatically determining femoral coordinate system, and also constructing SSM of the femoral bone. The coordinate system is exploited to align MR images to a base space of equal voxel size for the purpose of registration. Finally, principal component analysis (PCA) is applied on high dimensional data obtained from signed distances of the registered images. The implemented model is evaluated by reconstructing images, utilizing standard deviation (SD) ratio for each eigenvector obtained from the model.

In this paper, we propose a method of nursing-care text classification. We have proposed some nursing-care classification methods using fuzzy systems, standard three-layer neural networks, and support vector machines. Also we have proposed several types of feature vector definitions for expressing free style Japanese texts into numerical vectors. This paper proposes a novel feature vector definition and a support vector machine utilizing a decision tree (SVM-BDT) based classification system. From experimental results, the effectiveness of both feature definition and SVM-BDT-based classification system is shown.

© 2015. People are posting huge amounts of varied information on the Web as the popularity of social media continues to increase. The sentiment of a tweet posted on Twitter can reveal valuable information on the reputation of various targets both on the Web and in the real world. We propose a method to classify tweet sentiments by machine learning. In most cases, machine learning requires a significant amount of manually labeled data. Our method is different in that we use social bookmark data as training data for classifying tweets with URLs. In social bookmarks, comments are written using casual expressions, similar to tweets. Since tags in social bookmarks partly represent sentiment, they can be used as supervisory signals for learning. The proposed method moves beyond the basic "positive"/"negative" classification to classify impressions as "useful", "funny", "negative", and "other".

In this paper, a system of personal identification using data associated with corneal thickness is presented for ophthalmological patients. The data are measured by OCT (Optical Coherence Tomography). The proposed method equally divides the cornea into thirty-two fan-shaped segments, using thirty-two radiuses. It generates a couple of thirty-two-dimensional vectors (or a sixty-four-dimensional vector) for some test subject as the registered data, and adds it to a set. Each of the data consists of element values associated with minimum values and/or maximum values on the above radiuses. When the test subject takes medical practice, the proposed method generates two thirty-two-dimensional vectors (or a sixty-four-dimensional vector) for the corresponding test subject as the collation data. Then, the proposed method calculates the distance between the collation data with each of the vectors in the set. It judges the test subject corresponding to the given collation data to be that of the registered data with the shortest distance to the given data. Experimental results establish that the proposed method achieves 100 percent as the identification rate on assumption that the number of subjects is at most 45.

In our past works, a standard three-layer feedforward neural network based human activity estimation method has been proposed. The proposed method aims to record the subject activity automatically. The recorded data by MEMS-based monitoring devices include raw accelerometer data of his/her activity. From these data, we need to determine what the subject person was doing. In our conventional methods, some numerical datasets of accelerometer which are measured for every subject person were needed to train neural networks. In this paper, we propose an estimation method of subject behavior using fuzzy neural networks. The proposed fuzzy neural network based method can be trained by using fuzzy if-then rules which represent action primitives instead of numerical datasets from subject person.

In this paper, we present a method of selecting intraocular lens (IOL) power formulas for cataract patients. The method is based on support vector machines (SVM's) and genetic algorithm (GA). We assume that each of patients' data belongs to one of the classes named by three power formulas. We therefore consider the formula selection to be the issue of classifying patients' data. Since three formulas have variables associated with axial length and corneal refractive power, we always employ them as elements in data. If there are elements probably useful in classifying the data, they are found by GA. We construct the final discrimination model by SVM learning using data with the above elements. The model consists of three coordinate spaces, each of which has two regions corresponding to two formulas. A space provides a potential solution. We finally take a majority of the potential solutions, and determine the formula suitable to some patient specified by it. We show that the proposed method achieves the most favorable percentage of concordance for the classification, if the two-point crossover technique is adopted in GA.

In this paper, we propose a modified 4D macrophage simulator based on evolving cellular automata. Macrophages are well known in the biology area. Recently, we can visualize macrophages in vivo. Some observation data such as movies are available for analyzing macrophages' behavior. The goal of our research is to find rules of behavior of macrophages. As the first step for finding rules, we have developed a simulator which is able to imitate macrophages' behavior in four-dimensional space. In our previous research, just two short movies which were recorded behaviors of macrophages in a living mouse were available for constructing our simulator. Now, we have some more movies which are recoded more complex condition than the previous two movies. In our previous works, we have proposed an evolving 2D-cellular automata based method. Actually, macrophage observation data are multiple 2D images which are obtained by multiple depth. However, researchers desire to analyze macrophages in the 3D space with the target elapsed time. Therefore, we propose an evolving 3D-cellular automata based method in this paper. 2D cells which define neighborhoods are extended cubic cells in 3D space. Transition rules for cellular automata are defined in such cubic cells.

To prevent lifestyle diseases, this paper studies disease prediction using periodical health checkup data, daily monitoring to maintain healthy condition, and early life disease detection with medical imaging. To analyse periodical health checkup data, three approaches are introduced. The first approach is based on fuzzy set. It converts all attributes of health checkup data into fuzzy degrees by defining fuzzy membership functions. It enables us to manipulate all attributes in the same scale. The second approach analyses relationships between attributes of specific health examination data to cope with lifestyle diseases. It uses self-organizing maps, and clarifies the relationships among hemoglobin A1c (HbA1c), glutamic-oxaloacetic transaminase, glutamic-pyruvic transaminase, gamma-glutamyl transpeptidase, and triglyceride. The third approach predicts HbA1c fluctuations using decision tree. If we can predict the fluctuation, we can extract knowledge about what element will trigger developing diabetes. Through our examination, BMI will be the largest influencer about HbA1c fluctuations. Daily understanding of own condition is the first step of maintaining our health. A MEMS-based small and flexible monitoring device has been developed by the ERATO Maenaka human-sensing fusion project. We propose a condition estimation method using the monitoring device and FNN-based condition estimation. Experimental results show that it is a promising method for condition understanding. Cerebral vascular disease is one of major lifestyle diseases, and is caused by cerebral aneurysms. To predict the diseases, we should analyse cerebral arteries and aneurysms using magnetic resonance angiography images. This paper introduces an automated analysis method for early detection of aneurysms.

In this paper, we propose a method of nursingcare text classification. We have proposed some nursing-care classification methods using fuzzy systems, standard three-layer neural networks, and support vector machines. Also we have proposed several types of feature vector definitions for expressing free style Japanese texts into numerical vectors. This paper proposes a novel feature vector definition and a support vector machine utilizing a decision tree (SVM-BDT) based classification system. From experimental results, the effectiveness of both feature definition and SVM-BDT-based classification system is shown.

The vehicle routing problem(VRP) is an important issue in practical use. VRPs are one of combinatorial optimization problems. For solving such combinatorial problems, several evolutionary computation methods have been proposed. In practical situation, many complex constraint conditions and desired computation time are obstacles to use evolutionary computation methods to such problems. In this paper, an evolutionary computation based solver is developed. From some computational experiments for real tanker truck scheduling problem, effectiveness of the proposed method is shown.

Improving the quality of nursing care is crucial to maintaining the quality of life. Our objective is to develop a computer-aided evaluation system that enables nursing experts to improve the quality of nursing care. In our previous works, some classification systems based on fuzzy logic, neural networks, and SVMs were developed. Although a classification system with high performance for all nursing-care datasets is desirable, we focus on how to visualize the classification results in this paper. It is important to visualize the results for our nursing-care text classification system because the computer-aided system has to explain the reasons for obtaining such results to human experts. In this paper, a tree-type expression is considered for visualizing the classification results. To visualize classification results with the tree-type expression, we consider a decision tree technique. Word existence, dependency relations, and phrase-based feature vector definitions have been proposed in our previous works. In the present study, these three types of feature vector definitions are compared with one another from the viewpoint of understandability.

In the aging society such as Japan, it is very important to improve the quality of nursing-care for keeping our quality of life. Our final goal is to develop a computer aided evaluation system to improve the quality of nursing-care. For evaluating the quality of actual nursing, we have been collecting texts that are written by nurses using our Web based system. In our previous works, a SVM based classification system has been developed to classify such nursing-care texts, and a dependency relation based feature vector definition has also been proposed. The training data are pre-classified texts by a few nursing-care experts. Some texts in the training data are similar but classified into different classes. To classify the nursing-care texts with high accuracy, we need to tackle such ambiguous class labels in the training data. In this paper, we propose a k-nearest neighbor based classification system which can classify into classes with certainty grade.

【目的】本研究の目的は、臨床施設と大学が連携した看護研究支援システム構築への第一歩として、臨床看護師がどのようなテーマに関心を持ち看護研究を行っているかを明らかにすることである。【研究方法】日本看護系学会協議会の会員学会名簿より、特定の大学や病院が運営しているため入会制限のある学会を除き、学会会員数が1,000名以上である20学会が、2012年に発行した学術集会抄録集を分析の対象とした。対象より臨床看護師が筆頭研究者である発表演題を抽出し、KH Coderを用いて発表演題名のテキストマイニング分析を行った。【結果】対象4,289演題のうち臨床看護師が筆頭研究者である発表演題は、1,947件(45.4%)であり、その内の68.5%が研究者が臨床看護師のみの発表演題であった。テキストマイニングを用いた分析より、頻出語上位は、『患者』、『看護師』、『看護』であった。さらに、『患者』の共起語として<受ける><家族><検討><化学療法><支援>、『看護師』の共起語として<要因><新人><影響><手術室><認識>、『看護』の共起語として<患者><家族><終末期><がん患者><外来>が抽出された。研究者が臨床看護師のみの発表演題と、大学教員が含まれる発表演題とで頻出語を比較すると、前者では『検討』『取り組み』『効果』などの言葉が、後者では『要因』『支援』『影響』などの言葉が上位を占めた。【考察】臨床看護師は、数多くの看護研究を行い学会で発表していたが、先行研究によれば論文に至るものは少ない。臨床看護師は、患者家族、化学療法患者、新人看護師、手術室看護師、がん看護、終末期看護、外来看護に関するテーマで数多くの看護研究を行っていることが推測された。その内容は、日々の看護援助の取り組みの報告や直面している課題に対してであることが推察された。今後、臨床看護研究に大学が携わることでEvidenceの構築と看護の質の向上へと繋がる可能性が示唆された。(著者抄録)

In the aging society such as Japan, we feel large importance for improving the quality of nursing-care to keep our quality of life. Development of a computer aided evaluation system for improving the quality of nursing-care is our final goal. In order to evaluate the quality of actual nursing in wide areas in Japan, we have been collecting texts that are written by nurses using our Web based system. A SVM based classification system has been developed to classify such nursing-care texts, and a dependency relation based feature vector definition has also been proposed in our previous researches. When we train the SVM based classification system, pre-classified nursing-care texts by a few nursing-care experts are used as a training data set. Some texts in the training data are similar but classified into different classes. To classify the nursing-care texts with high accuracy, we need to extract numerical features that can express characteristics of the original text. In this paper, we explain some feature vector definitions and propose a directed graph based feature vector definition.

The dynamics behaviour and underlying principles of macrophages movement within living bodies still being unknown, we are currently studying unique sets of pictures in order to better understand them and develop a simulator able to successfully reproduce them. In previous works, we presented our early efforts in building a simple 2D macrophages simulator, as well as a new set of 3D data and a new tool to visualize them. In this paper, our goal is to progress towards a full analysis of our 3D data by extending our visualizing tool into an analysing tool, able to fully break the data about macrophages movement. By applying object recognition and object tracking techniques, we developed a tool able to display and record all information relative to the moves of a given three-dimensional macrophages set, information we will be able to use in order to develop a new 3D macrophages simulator.

We proposed a standard three-layer feedforward neural network based human activity estimation method. The purpose of the proposed method is to record the subject activity automatically. Here, the recorded activity includes not only actual accelerometer data but also rough description of his/her activity. In order to train the neural networks, we needed to prepare numerical datasets of accelerometer which are measured for every subject person. In this paper, we propose a fuzzy neural network based method for recording the subject activity. The proposed fuzzy neural network can handle both real and fuzzy numbers as inputs and outputs. Since the proposed method can handle fuzzy numbers, the training dataset can contain some general rules, for example, "If x and y axis accelerometer outputs are almost zero and z axis accelerometer output is equal to acceleration of gravity then the subject person is standing."

In the aging society such as Japan, it is very important to improve the quality of nursing-care for keeping our quality of life. Our final goal is to develop a computer aided evaluation system to improve the quality of nursing-care. For evaluating the quality of actual nursing, we have been collecting texts that are written by nurses using our Web based system. In our previous works, a SVM based classification system has been developed to classify such nursing-care texts, and a dependency relation based feature vector definition has also been proposed. The training data are pre-classified texts by a few nursing-care experts. Some texts in the training data are similar but classified into different classes. To classify the nursing-care texts with high accuracy, we need to tackle such ambiguous class labels in the training data. In this paper, we propose a k-nearest neighbor based classification system which can classify into classes with certainty grade.

Macrophages are well known cells in biology. While macrophage plays an important role in the immune system, the behavior rules of macrophages are unknown. In our research, the final goal is to find action rules for behavior of macro phages. Our current aim is to develop a simulator which can reproduce the movement of macro phages. As the first step, we try to develop a simulator that can imitate macrophages' behavior in three-dimensional space. In order to develop a simulator, some movies of actual macrophages in live mice are used. After parsing the movies, we try to develop a simulator which can imitate behaviors as same results as the observed. Our proposed simulator is based on cellular automata which are evolved by genetic algorithms. From the results of our simulator, we can see that the artificial macrophages behave as the similar movement of actual macrophages.

© 2014 TSI Press. For maintaining our daily healthcare, we need to understand our own physical condition. In understanding such data, additional information such as what the subject is doing at that time is needed. For example, let us assume that we have a record of a certain heart rate 90. If such value is observed when the subject person was sleeping, that value is high and the subject may have some trouble on his/her health. On the other hand, when the subject was running, the subject has no problem on his/her health. In this paper, we propose a combined system for maintaining our healthcare. Our proposed system consists of both systems; (1) a fuzzified neural network based unusual condition detection and (2) a standard neural network based action estimation. The proposed system can handle multiple kinds of sensors&#039; data. In this paper, the following three kinds of sensors were handled; (1) three-axis acceleration data, (2) heart rate, and (3) breathing rate. From experimental results, the effectiveness of our proposed system is shown for understanding our conditions.

Continuous human monitoring will be substantially useful to realize a high quality of life society. In the previous work, we fabricated a prototype system for monitoring an electrocardiograph, heart rate (HR), 3 axes human body acceleration and temperature for human body and human circumstances, simultaneously. These data are transmitted to the host PC and analyzed for the human activities or conditions recognitions. Above all a heart rate variability (HRV) that calculated from HR is extremely valuable for recognizing a mental or physical stress of human subjects. In this study, we demonstrate a fuzzy logic HR extraction algorithm on the daisy chain shaped wearable prototype device to realize an autonomous HRV monitoring system. On-board fuzzy logic algorithm not only reduces the communication traffic but also improves an accuracy of the HR extraction comparing to the simple threshold methods. © 2013, TSI® Press.

It is very important for us to improve the nursing-care quality. To improve the nursing-care quality, the "Web based Nursing-care Quality Improvement System" have been proposed and operating continuously. In the Web-based system, freestyle Japanese texts, which are called "nursing-care texts," are collected through the Internet in Japan for evaluating actual nursing-care process. The nursing-care experts can evaluate actual nursing-care process by reading the collected nursing-care texts carefully and then recommend some improvements to nurses. However, it is hard to do the above-mentioned evaluation process for a large number of nurses because the number of nursing-care experts who can evaluate the nursing-care texts is a few. In order to assist nursing-care experts in evaluating the nursing-care texts, a computer aided nursing-care text classification system has been developed. In this paper, we propose a novel feature definition for the nursing-care text classification system. Dependency relations between terms are extracted from the nursing-care texts. The extracted dependency relations are used as feature values that represent characteristics of the nursing-care text. © 2013 Copyright TSI® Press.

In chemical plants, a process control system (PCS), which is composed of instruments and a distributed control system(DCS), is widely used for the plant automation. Important infrastructure equipment is considered critical to the safe operation of a plant. Failures of the PCS can lead to extremely dangerous accidents such as leakage of poisonous gas or fluid potentially leading to an explosion. Therefore, the plant maintenance department conducts planned activities to prevent failure. However, it is very difficult in actuality to reduce the failure to zero by only planned maintenance. Consequently, to detect a sign of failure early, the plant operators carefully monitor the critical process variables and patrol the facilities in the field as part of their work. Additionlly, nowadays advanced present chemical plants in automation increase the operator's role such as ensuring safety, saving energy and environment not just controlling amount of production as planned. As the result, their burden of work is higher than before. In this paper, we propose a fault detection method of the instrument devices based on fuzzified neural networks with less adjusting parameters, which helps the operator's detection.

In order to understand our physical condition, we need to record the detail of physical condition data like the heart rate. However, for understanding such data, additional information such as what the subject is doing at that time is needed. We propose a combined system which consists of a fuzzified neural network based unusual condition detection and a standard neural network based action estimation. From experimental results, the effectiveness of our proposed system is shown for understanding our conditions.

Since Japan is one of the most aging countries, it is very important for us to improve the nursing-care quality. For improving the nursing-care quality, a Web-based nursing-care quality improvement system have been proposed and operating experimentally and continuously. A kind of collected data by the Web-based system is freestyle Japanese text called "nursing-care texts". The nursing-care texts are used for evaluating actual nursing-care process. In order to assist nursing-care experts in evaluating the nursing-care texts, a computer aided nursing-care text classification system has been developed. In this paper, we propose a phrase based feature vector definition for classifying the nursing-care texts. The dependency relation based feature vector definition has been proposed in our previous work. As another feature vector definition method, we propose a phrase based feature vector definition method. Phrases are found by using the dependency relation analysis and stored into a phrase list. We also define a similarity between phrases because each phrase consists of some kinds of words. From experimental results, we show that our phrase based feature vector contributes the classification performance.

Macrophages are an essential part of our immune system and one of the well-known cells in biology. However, despite its importance its usual behaviour and moving rules are mostly unknown due to the lack of observatory data for the macrophages. An experiment was conducted in an attempt to obtain new data by observing the macrophages in a living rat and produced a couple of movies. Our objective is to exploit these results to develop a simulator that is able to reproduce the moves of a macrophage population. After the movie analysis, we identified a few characteristics to implement into a first simple simulator. Although its performance is very limited, it represents a first step for further research of understanding macrophages.

For monitoring and estimating our daily activity, some kinds of devices are available. One of such kinds of monitoring devices is a MEMS based prototype which is developed by the Maenaka Human Sensing Fusion Project. We have developed a estimation method of human activity from three-axis acceleration data using the above-mentioned prototype. This method can estimate our unit activities, such as (1) walking, (2) running, (3) sitting, (4) lying, and (5) standing. In this paper, we propose a system that can find unusual situation from ECG data. Our proposed system is based on the fuzzified neural networks. The fuzzified neural network is trained by using sensing data with reliability grade. Since the fuzzified neural network learns normal state of the subject person, we can understand the ECG state of the subject when we analyze fuzzy outputs from the trained fuzzified neural network. This paper shows estimation results by using actual monitoring data which contains normal state, and artificial unusual data. From the results for the actual monitoring data, we can see that our proposed system was able to estimate the testing data as normal. From the results of estimating artificial unusual data, our proposed system can find the subject person's unusual situation. (C) 2013 The Authors. Published by Elsevier B.V.

In this paper, we propose rarity of a Web page in a category given by a user to find useful information that a few people know. A rare Web page is a page that belongs to a given category and that is atypical in the category. We define a probability that the page is a rare Web page in the given category as a rarity score. The rarity score is a product of a relevancy score and an atypicality score. The relevancy is a probability that a Web page belongs to a category given by a user. The atypicality is a conditional probability that a page is atypical in the category when it belongs to the category. Both probabilities are calculated by using tags of social bookmark services and words in Web pages. We evaluated the proposed relevancy score by classifying whetherWeb pages belong to a certain category. We also evaluated the proposed rarity as a metric for ranking Web pages, and compared the rankings by relevancy and atypicality. We confirmed usefulness of the rarity score to find relevant and atypical pages.

We have proposed a fuzzified neural network based failure diagnosis for chemical plants. The proposed method used the fuzzified neural network which is a kind of standard layered neural networks with fuzzy weights and fuzzy biases. For constructing fuzzy models, time series data from process control systems is modified and used as the training data. In order to use the raw time series data as the training data, a reliability grade is assigned to each input-output pair of the raw time series data. Since the training data plays significant role for the suitability, it is important for our method to assign a reliability grade to every input-output pair of the raw time series data. In this paper, we propose a novel grade assignment method for the fuzzified neural network based failure diagnosis.

In order to maintain our healthcare, daily monitoring our physical condition is very important. A multiple microelectromechanical system (MEMS) based monitoring system has been developed. The MEMS based monitoring system is developed for aiming noninvasive and unconstrained monitoring to our physical condition. From a viewpoint of the hardware, the smaller size of such monitoring system is better for us. Additionally, the smaller size is better for power consumption. Therefore, we need to design and develop a very small size of monitoring software. Two-step abstraction method for estimating human behavior had been proposed. In this paper, we propose an extended method of the two-step abstraction for ultra-small hardwares. © 2012 TSI Press.

Continuous human monitoring is substantially useful to realize a high QoL (quality of life) society. In the previous work, we fabricated a prototype system for monitoring an electrocardiograph (ECG), heart rate (HR), 3 axes human body acceleration, temperature for human body and human circumstances, simultaneously. These data are transmitted to the host PC and used for analyzing the human activities and conditions such as a heart rate variability (HRV). The HRV that calculated from HR is valuable for recognizing a mental or physical stress of human subjects. In this study, we demonstrate a fuzzy logic HR extraction algorithm on the prototype system to realize an autonomous HRV monitoring system. On-board fuzzy algorithm will reduce the communication traffic and improve an accuracy of the HR extraction. From the implementation result, the error ratio of the HR extraction is improved from 0.9 % to 0.4 %. © 2012 TSI Press.

Recently, "Web based Nursing-care Quality Improvement System" have been proposed and operating continuously for improving the nursing-care quality in Japan. For evaluating actual nursing-care process, freestyle Japanese texts which are called "nursing-care texts" are collected through the Internet. The nursing-care experts read the collected nursing-care texts carefully to evaluate actual nursing-care process. Then they make a recommendation which includes some improvements, and send it to each nurse. The number of nursing-care experts who can evaluate the nursing-care texts is a few. Hence, it is hard to perform the above mentioned evaluation process because of a large number of nurses. In order to assist nursing-care experts in evaluating the nursing-care texts, we have been developing a computer aided nursing-care text classification system. In this paper, first, we introduce our computer aided nursing-care text classification system. Then we propose a method to improve the classification performance of the nursing-care text classification system. In our proposed method, dependency relation between terms is extracted from the nursing-care text. The extracted dependency is used as a feature value which represents characteristics of each nursing-care text. From some experimental results for the actual nursing-care text sets, we show that our proposed feature definition is effective for improving the classification performance.

For maintaining our health, we need to know our physical condition. The daily monitoring which is one of methods for taking our physical condition is very important. Because we have to attach such a monitoring device on some part of our body for taking our condition, noninvasive monitoring devices are needed. For realizing such a noninvasive monitoring device, several micro electro mechanical system (MEMS) based monitoring system has been developed. From a viewpoint of attaching monitoring devices on our body, the smaller size of such monitoring system is better. An ultra-small monitoring device should have ultra-small size power source. Therefore, we need to design and develop a very small size of monitoring software with lower power consumption.