JP2008264352A - Menstrual cycle estimation device and menstrual cycle estimation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a menstrual cycle estimation device for women and a menstrual cycle estimation method for women, which are capable of clearly grasping two-phase characteristics of a body temperature in a menstrual cycle of a subject on the basis of detection data of the body temperature over several days. <P>SOLUTION: A control unit 5 of a sensor device 1 transfers body temperature detection data D1 detected while a subject H0 is asleep by using a body temperature detection part 3 to an outside processor 11. The processor 11 time serially links the body temperature detection data D1 for several days to generate body temperature data sequences O and uses the body temperature data sequences O to construct a hidden Markov model HMM having two hidden states corresponding to a low temperature phase q<SB>1</SB>and a high temperature phase q<SB>2</SB>in the menstrual cycle. Thus, the processor 11 operates maximally likely state sequences Q<SP>*</SP>in which the hidden Markov model outputs the body temperature data sequences O and, on the basis of the state sequences Q<SP>*</SP>, estimates the two-phase characteristics of the body temperature of the subject H0 for several days when the body temperature has been measured. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  In the present invention, for example, whether the subject is in the high temperature period or the low temperature period based on the body temperature (including the body surface temperature and the deep temperature, and so on) measured over a plurality of days. The present invention relates to a menstrual cycle estimation apparatus and a menstrual cycle estimation method.

  Conventionally, in order to understand the hormone balance status of women for the purpose of knowing the cycle that is a standard for pregnancy and contraception, daily fluctuations in basal body temperature are tracked over several months, and the fluctuation cycle and fluctuation range are numerical values. Analytical methods have been used. Although it is desirable to measure the body temperature during sleep, which has the least metabolism during the day, it is generally difficult to measure body temperature during sleep. For this reason, for example, the mouth temperature at the time of getting up is measured as a body temperature, and the measured body temperature has been used instead of the basal body temperature (see, for example, Patent Document 1). However, in view of the diversification of women's lives, it is difficult to make the measurement time, measurement method, and other conditions constant every morning. Continuing measurement over a long period of time is inconvenient and requires a short mouth temperature. In measurement, artifacts (distortion of measurement signals due to external factors) were likely to occur in the measurement results. For this reason, a sensor device for measuring the temperature of the body surface (body surface temperature) is attached to the abdomen of a person who is sleeping, etc., and the basis is based on the body surface temperature automatically measured over a long period of time by this sensor device. What determines the representative temperature of the day corresponding to body temperature is known (for example, refer patent document 2). At this time, the representative temperature was, for example, the highest temperature or the average temperature of the measured body surface temperature.

JP-A-11-316161 JP 2004-163391 A

  By the way, in the prior art, the close contact state of the sensor device with the body surface may change due to the measurement subject's turning over, and the measurement results may vary. For example, in a state where the sensor device is attached to the abdomen of the measurement subject, the closeness of the sensor device is low when lying on its back, whereas the closeness of the sensor device is high when lying on its back. Furthermore, if you lie on your back, there is a possibility of a rise in body surface temperature due to a temporary increase in blood flow due to abdominal stimulation, or a decrease in body surface temperature due to a decrease in blood flow due to abdominal pressure. In addition, there is also a possibility of a rise in body surface temperature associated with the use of an electric blanket. As a result, artifacts are included in the measurement result of the body surface temperature, and the measurement accuracy of the body surface temperature and the representative temperature tends to decrease.

  In the prior art, the representative temperature is determined in the range of the measured body surface temperature. At this time, even in the prior art, it is possible to some extent to grasp the fluctuation of the basal body temperature using the representative temperature. However, since the body surface temperature is generally susceptible to the influence of outside air, blood flow, etc., if the maximum or average measured temperature is determined as the representative temperature, the representative temperature may cause artifacts due to the influence of the outside air, etc. There is. Here, in grasping the biphasic basal body temperature composed of the high temperature period and the low temperature period, it is necessary to accurately grasp the + 0.3 ° C. rise from the low temperature period to the high temperature period. On the other hand, in the prior art, there is a problem that it is difficult to distinguish between an increase in body temperature in a high temperature period or a change due to measurement artifacts.

  Thus, since the representative temperature is affected by measurement artifacts, it has been difficult to determine whether the measurement subject is in the high temperature period or the low temperature period based on the representative temperature. As a result, the biphasicity of basal body temperature and the period were easily unclear, and it was not possible to easily grasp the hormone balance status of women.

  The present invention has been made in view of the above-described problems of the prior art, and an object of the present invention is to provide a menstruation that can clearly grasp the biphasicity of the body temperature of a person to be measured based on measurement data of body temperature over a plurality of days. The object is to provide a cycle estimation device and a menstrual cycle estimation method.

  In order to solve the above-described problem, the invention of claim 1 is based on the body temperature data series obtained by measuring the body temperature of the subject over a plurality of days, and the subject is in either the high temperature period or the low temperature period. A menstrual cycle estimation device that estimates whether there is a hidden Markov model having two hidden states of a high-temperature phase and a low-temperature phase using the body temperature data sequence, and the hidden Markov model uses the body temperature data sequence The most likely state sequence to be output is obtained, and the state sequence is used to identify the body temperature biphasicity in the menstrual cycle.

  In the invention of claim 2, a casing that can be always attached to the body of the measurement subject, body temperature detection means that is provided on the body surface side of the measurement subject in the casing, and detects the body temperature, and ends from the start time every measurement day Reading means for reading body temperature detection data by the body temperature detection means at a predetermined time interval until the time, storage means for storing body temperature detection data by the reading means, and stored in the storage means Data series generation means for generating body temperature data series by connecting body temperature detection data over a plurality of days, model construction means for constructing the hidden Markov model using the body temperature data series by the data series generation means, A state sequence calculation unit for obtaining a maximum likelihood state sequence in which the hidden Markov model by the model construction unit outputs the body temperature data sequence; Using the state sequence has a configuration and a state specifying means for specifying a biphasic temperature in the menstrual cycle of the subject.

  According to a third aspect of the present invention, when the state specifying unit performs a filtering process using a median filter on the state series and measures each body temperature detection data for a state included in the state series after the filtering process. It is set as the structure made into the state of a to-be-measured person.

  According to a fourth aspect of the present invention, the data series generation means deletes values other than the preset temperature and the allowable temperature change range from the body temperature detection data, and connects the remaining body temperature detection data to form a body temperature data series. Out-of-range data deleting means for generating the pre-processing means for performing filter processing using a median filter on the body temperature data series generated using the out-of-range data deleting means, the model construction means The hidden Markov model is constructed using the body temperature data series after the filter processing by the preprocessing means.

  In the invention of claim 5, the model construction means constructs the hidden Markov model in a state where the body temperature detection data of the body temperature data series has discrete values with respect to the temperature.

  In the invention of claim 6, the model construction means constructs the hidden Markov model in a state where the body temperature detection data of the body temperature data series has a continuous value with respect to the temperature.

  According to a seventh aspect of the present invention, the model construction means is constituted by an external processing device, and the casing is provided with transfer means for transferring body temperature detection data stored in the storage means to the external processing device. .

  In the invention of claim 8, the menstrual cycle estimation for estimating whether the subject is in the high temperature period or the low temperature period using a body temperature data series obtained by measuring the body temperature of the subject over a plurality of days. A method for constructing a hidden Markov model having two states, a high temperature phase and a low temperature phase, using the body temperature data sequence, and obtaining a maximum likelihood state sequence from which the hidden Markov model outputs the body temperature data sequence. The state temperature is used to identify the body temperature biphasic during the menstrual cycle.

  In the invention of claim 9, the menstrual cycle for estimating which state the polymorphism of the body temperature in the menstrual cycle is based on the body temperature data series obtained by measuring the body temperature of the subject over a plurality of days A method for estimating a hidden Markov model having a plurality of hidden states corresponding to polymorphism of body temperature using the body temperature data series, wherein the hidden Markov model outputs the body temperature data series A series is obtained, and the polymorphism of body temperature in the menstrual cycle of the measurement subject is specified using the state series.

  According to the first and eighth aspects of the invention, a hidden Markov model having two states of a high temperature phase and a low temperature phase is constructed using the body temperature data series, and the hidden Markov model outputs the body temperature data series. Since the state series is obtained, the state of the person to be measured over a plurality of days when the body temperature data series is measured can be specified by determining the state included in the state series. As a result, even when measurement artifacts occur in the body temperature data series due to the influence of outside air or the like, it is possible to clearly grasp the biphasicity of the measurement subject.

  According to the invention of claim 2, since the body temperature detecting means is provided on the body surface side of the measurement subject in the casing, the reading means is provided at predetermined time intervals from the start time to the end time. Body temperature detection data is read using the body temperature detection means, and the storage means stores the body temperature detection data by the reading means. And since the data series generation means generates body temperature data series by connecting the body temperature detection data stored in the storage means for a plurality of days, the model construction means uses the body temperature data series by the data series generation means. A model parameter is calculated, and a hidden Markov model composed of the model parameter is constructed. Thereby, the state series calculation means can obtain the maximum likelihood state series from which the hidden Markov model formed by the model construction means outputs the body temperature data series. For this reason, the state specifying means can specify the state of the measurement subject when each body temperature detection data is measured using the state series by the state series calculating means.

  According to the invention of claim 3, since the state specifying means performs the process using the median filter for the state series, it is possible to remove the discontinuous detachment state included in the state series. As a result, the state does not repeatedly switch in units of hours or days, and the biphasicity of the subject who repeats the high temperature period and the low temperature period with a periodicity of about one month can be clearly grasped.

  According to the invention of claim 4, the out-of-range data deleting means deletes values other than the preset temperature and the allowable range of temperature change from the body temperature detection data as abnormal values, and consists of a value within the allowable range. The body temperature data series is generated by concatenating the remaining body temperature detection data. Generally, since the body temperature measurement range of the human body is about 32 ° C. to 40 ° C., it is considered that the temperature detected by the body temperature detection means in a thermal equilibrium state does not deviate from the allowable range of 32 ° C. to 40 ° C., for example. In the case where the body temperature detection data has a value outside this temperature range, for example, it may be considered that the person to be measured removes the casing from the body or receives external heat.

  Further, since the body temperature of the measurement subject is relatively stable during sleep, it is considered that the measurement temperature does not change beyond the allowable range of ± 1 ° C. at a measurement interval of about 10 minutes, for example. In the case where the body temperature detection data has a value outside the allowable range of the temperature change, for example, when the person to be measured temporarily wakes up or when the person to be measured jumps off the bedding.

  Thus, it is considered that the body temperature detection data other than the allowable temperature range or the allowable temperature change range has a low contribution rate to the state of the subject. In addition, if a hidden Markov model is constructed using body temperature detection data other than the allowable range of temperature and temperature change, the state estimation accuracy may be reduced. For this reason, the out-of-range data deleting means deletes values outside the allowable range from the body temperature detection data and stores the body temperature data series.

  Further, since the preprocessing means performs a filtering process using the median value filter on the body temperature data series stored by the out-of-range data deletion means, it is possible to remove discontinuous outliers included in the body temperature data series. it can. And since the model construction means constructs the hidden Markov model using the body temperature data series after the filtering process by the preprocessing means, the hidden Markov model is compared with the case where the body temperature data series including outliers is used. It is possible to prevent the outlier from affecting the model parameters of the model, and it is possible to improve the estimation accuracy of the state of the person being measured.

  According to the invention of claim 5, the model construction means constructs the hidden Markov model in a state where the body temperature detection data of the body temperature data series has a discrete value with respect to the temperature. Therefore, the hidden Markov model is a discrete type. It has a body temperature detection data output probability consisting of a discrete probability distribution according to the body temperature detection data. For this reason, a state series can be obtained using a body temperature data series composed of body temperature detection data discrete with respect to temperature and a discrete hidden Markov model. In addition, since the discrete hidden Markov model is used, the time, storage capacity, and the like required for the model construction process can be reduced as compared with the case where the continuous hidden Markov model is used.

  According to the invention of claim 6, since the model construction means constructs the hidden Markov model in a state where the body temperature detection data of the body temperature data series has a continuous value with respect to the temperature, the hidden Markov model is a continuous type. It has a body temperature detection data output probability consisting of a continuous probability distribution according to the body temperature detection data. For this reason, a state series can be obtained using a body temperature data series composed of body temperature detection data continuous with respect to temperature and a continuous hidden Markov model. Further, since the continuous hidden Markov model is used, the accuracy of the state sequence can be improved as compared with the case where the discrete hidden Markov model is used.

  According to the invention of claim 7, the model construction means is constituted by an external processing device, and the casing is provided with transfer means for transferring the data stored in the storage means to the external processing device. Here, in order to estimate the state of the person to be measured using the body temperature data series measured over a plurality of days and grasp the fluctuation period of the basal body temperature of the woman (the person to be measured), the body temperature for one cycle or more It is preferable to construct a hidden Markov model using the detected data. In this case, the hidden Markov model is better when using an external processing device than when using a microcomputer or a mobile phone that forms a control circuit in a small casing that is inserted into the subject. Is efficient to build.

  Therefore, in the present invention, body temperature detection data stored in the storage means is transferred to an external processing device such as a server computer using a transfer means such as a wired system or a wireless system. For this reason, the processing device can process a large amount of body temperature detection data, and can quickly build a hidden Markov model.

  According to the invention of claim 9, a hidden Markov model having a plurality of hidden states corresponding to body temperature polymorphism is constructed using a body temperature data series, and the hidden Markov model outputs a body temperature data series. Since the state series is obtained, by determining the state included in the state series, it is possible to specify the state having the polymorphism of the measurement subject over a plurality of days when the body temperature data series is measured. As a result, even when measurement artifacts occur in the body temperature data series due to the influence of outside air or the like, it is possible to clearly grasp the polymorphism of the measurement subject. Further, for example, when an intermediate temperature phase is provided between the low temperature phase and the high temperature phase, the rising change point from the low temperature phase to the high temperature phase and the falling change point from the high temperature phase to the low temperature phase can be examined in detail.

  Hereinafter, a menstrual cycle estimation apparatus according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

  First, FIG. 1 to FIG. 10 show a first embodiment, in which 1 is a wearable sensor device (hereinafter referred to as sensor device 1) that constitutes a menstrual cycle estimation device together with an external processing device 11 to be described later. Thus, the sensor device 1 is roughly constituted by a casing 2, a body temperature detection unit 3, a control unit 5, a display unit 9 and the like which will be described later.

  Reference numeral 2 denotes a casing that forms the main body (main support) of the sensor device 1, and the casing 2 is formed in a substantially oval (oval) box shape using, for example, a resin material as shown in FIGS. 1 to 4. The circuit board 4 and the like to be described later are accommodated in the interior. Further, an opening 2A penetrating in a substantially circular shape is provided on the body surface side (front side) of the subject H0 in the casing 2. Here, the opening 2 </ b> A is located on the center side of the casing 2.

  Reference numeral 3 denotes a body temperature detection unit (body temperature detection means) that detects body temperature, and the body temperature detection unit 3 is attached to an opening 2A located on the center side of the casing 2 as shown in FIG. Moreover, the body temperature detection part 3 is comprised by the temperature measuring element which consists of a thermistor etc., and the metal cover which covers this temperature measuring element, for example. And since the body temperature detection part 3 is arrange | positioned in the casing 2 at the body surface side of the to-be-measured person H0, a cover contacts the body surface of the to-be-measured person H0, and body temperature conducts heat to a temperature measuring element through a cover. Thereby, the body temperature detection part 3 outputs the signal according to the body temperature of the to-be-measured person H0 by the temperature sensor.

  The body temperature detector 3 is configured to directly detect the body temperature by contacting the body surface of the person H0 to be measured. However, the present invention is not limited to this. For example, when an undergarment or the like is sandwiched between the body temperature detecting unit 3 and the body surface of the person to be measured H0, the body temperature detecting unit 3 is connected to the person to be measured H0 via the underwear or the like. It is good also as a structure which detects indirectly the body temperature. That is, the body temperature detection unit 3 does not need to detect the absolute body temperature of the person to be measured H0, and only needs to detect a relative temperature that changes according to the body temperature. Therefore, if the daily measurement conditions are substantially constant, for example, the body temperature detection unit 3 may be in close contact with the body surface of the person to be measured H0 with one piece of underwear interposed therebetween. The body temperature detection unit 3 is not limited to the one that detects the temperature of the body surface of the person to be measured H0, and may be configured to use, for example, a deep thermometer that detects the temperature inside the body of the person to be measured H0.

  Reference numeral 4 denotes a circuit board accommodated in the casing 2, and a control unit 5 as a reading means composed of a microcomputer or the like is mounted on the circuit board 4 as shown in FIG. 5. The control unit 5 has an input side connected to the body temperature detection unit 3 and an output side connected to a display unit 9 described later. Further, the control unit 5 is provided with a storage unit 6 composed of, for example, a ROM, a RAM or the like as a storage means.

  Here, the storage unit 6 stores in advance a program for operating the control unit 5, a start time ts, an end time te, and a time interval Δt used in the program, and a body temperature T 1 described later by the operation of the control unit 5. Is stored.

  At this time, the start time ts and the end time te are set as, for example, 0:00 am (ts = 0: 0 am) and 6:00 am (te = 6: 00 am) as sleeping time, and the time interval Δt is, for example, 10 It is set to a value of about minutes. Note that the start time ts and the end time te are set to the start time and the end time of bedtime if the person to be measured H0 is a night worker, for example. Also, the time interval Δt is not limited to 10 minutes, and is appropriately set according to the measurement conditions and the like, for example, between about 1 to 30 minutes.

  The control unit 5 has a timer 7 for measuring time, and is connected to a switch unit 8 including, for example, two button switches 8A and 8B. The control unit 5 is driven by a power source 5A such as a coin-type lithium battery mounted on the casing 2 and operates by reading a program from the storage unit 6 by operating the switch unit 8. Thereby, when the time by the timer 7 reaches the start time ts, the control unit 5 responds to the detected temperature from the body temperature detection unit 3 at regular time intervals Δt from the start time ts to the end time te. The body temperature detection data D1 is read.

  Then, the control unit 5 sequentially stores the body temperature detection data D1 in the storage unit 6. As a result, the storage unit 6 stores a plurality of body temperature detection data D1 from the start time ts to the end time te.

  The storage unit 6 has a value within a preset allowable temperature range (for example, 32 ° C. to 40 ° C.) in the body temperature detection data D 1 by the control unit 5 and an allowable temperature change range (for example, ± 1 ° C. / The body temperature detection data D1 is stored with a value within 10 minutes) as a normal value and a value outside the allowable range as an abnormal value. Specifically, the body temperature detection data D1 is stored as it is as a normal value, and missing data (for example, an error flag that can be distinguished from a normal value) is stored as an abnormal value.

  Here, since the body temperature measurement range of the human body is generally about 32 ° C. to 40 ° C., the allowable temperature range is that the temperature detected by the body temperature detector 3 in a thermal equilibrium state deviates from the range of, for example, 32 ° C. to 40 ° C. It is decided based on the idea that there is nothing to do. In addition, the body temperature of the measurement subject is relatively stable during sleep. For this reason, the allowable range of temperature change is determined based on the idea that it does not change beyond the allowable range of ± 1 ° C., for example, before and after the time interval Δt of about 10 minutes.

  The allowable temperature range and the allowable temperature change range are not limited to the exemplified values, and may be set as appropriate in consideration of measurement conditions and the like.

  The body temperature detection data D1 may be the detected body temperature T1 itself, for example, a difference value between a preset reference temperature numerical value T0 and body temperature T1. In this case, the reference temperature value T0 is preferably set to, for example, an average value (for example, 34 ° C.) of the body temperature T1 in order to reduce the data capacity.

  Reference numeral 9 denotes a display unit comprising a liquid crystal screen or the like provided on the back side of the casing 2, and the display unit 9 is connected to the control unit 5 and displays, for example, the driving state of the control unit 5. The display unit 9 constitutes transfer means for transferring the body temperature detection data D1 stored in the storage unit 6 to an external processing device 11 to be described later. By operating the switch unit 8, for example, a QR code (registered trademark) is provided. Are displayed. Here, the two-dimensional code includes information on the destination address such as the body temperature detection data D1 for one day (one night) stored in the storage unit 6 and the URL (Uniform Resource Locator) of the data. For this reason, the person to be measured H0 reads the information in the two-dimensional code into the cellular phone PT by using the cellular phone PT or the like having a two-dimensional code reading function. As a result, as shown in FIG. 6, the person under test H0 can access the mobile phone PT to the external processing device 11 via the Internet or the like and transfer the body temperature detection data D1 to the processing device 11. .

  Further, the person to be measured H0 may input the menstrual start date data by operating the switch unit 8 of the sensor device 1. In this case, the person to be measured H0 transfers the menstrual start date data to the processing device 11 by performing the two-dimensional code display, reading, and transmission operations using the sensor device 1 and the cellular phone PT as described above. Can do.

  The transfer means uses a two-dimensional code display by the display unit 9. However, the present invention is not limited to this, and a connection portion such as a connector for connecting the control unit 5 and the cellular phone PT using, for example, a wired method using cable connection or a wireless method using infrared or Bluetooth is provided. The transfer means may be configured as described above.

  Reference numeral 10 denotes a clip as a mounting means for attaching the casing 2 to the person to be measured H0. The clip 10 is located on the center side of the casing 2 and has a flexible band shape extending toward the upper part (outside). The tip is configured to be able to hold clothes or the like. Therefore, as shown in FIG. 1, the clip 10 fixes the sensor device 1 to the abdomen of the person to be measured H0 by sandwiching underwear such as shorts.

  The sensor device 1 is not limited to the abdomen of the person to be measured H0, and may be configured to be fixed to the chest of the person to be measured H0, for example. In this case, the clip 10 is attached to an underwear or the like worn at bedtime.

  Reference numeral 11 denotes a processing device that estimates whether the person to be measured H0 is in a high temperature period or a low temperature period using body temperature detection data D1 measured by the sensor device 1 as shown in FIG. 11 constitutes a data series generation means, a model construction means, a state series calculation means, and a state identification means. The processing device 11 is configured by, for example, a server computer and has a storage circuit including a ROM, a RAM, and the like.

Here, as shown in FIG. 7, the processing apparatus 11 includes a hidden Markov model HMM (hereinafter referred to as a model HMM) having _two states Q = {q ₁ , q ₂ } of a low temperature phase q ₁ and a high temperature phase q _2. ). At this time, the model HMM has a state transition probability matrix A = {a _ij } (probability of transition from state i to state j), body temperature detection data output probability matrix B = {b _j (k)} (b _j (k) Has the model parameter λ = {A, B, π} of the initial state probability matrix π = {π _i } (probability of state i in the initial state).

  Then, as shown in FIG. 8, the processing device 11 performs a menstrual cycle estimation process using the body temperature detection data D1 and the model HMM, and for a plurality of days (for example, 6 months) when the body temperature detection data D1 is measured. It is estimated whether the measurer H0 is in a high temperature period or a low temperature period.

The outline of a specific procedure of the menstrual cycle estimation process is as follows. First, the processing device 11 accumulates body temperature detection data D1 of the measurement subject H0 transferred from the sensor device 1 over a plurality of days (for example, 6 months), and connects these body temperature detection data D1 to body temperature data. A sequence O = {O ₁ , O ₂ ,..., O _T } is generated. Next, the processing apparatus 11 constructs a model HMM using this body temperature data series O.

  Here, when the model HMM is constructed, the model parameter λ = {A, B, π} of the model HMM is calculated using the body temperature data series O. At this time, as will be described later, the model parameter λ can be obtained using, for example, a Baum-Welch learning algorithm (Baum-Welch algorithm). Thereby, the processing apparatus 11 can construct | assemble the model HMM from which the likelihood of the body temperature data series O becomes the maximum.

The model HMM is constructed using body temperature detection data O _t having discrete values with respect to temperature, as will be described later. Therefore, the body temperature detection data output probability matrix B of the model HMM is configured by a discrete probability distribution according to the discrete body temperature detection data O _t .

Next, the processing device 11 may use the model HMM constructed using the body temperature data series O, and the model HMM may output the maximum likelihood path through which the body temperature data series O is output, that is, the body temperature data series O. Find the sequence Q ^* of the highest state change. At this time, the maximum likelihood state sequence Q ^* can be obtained using, for example, a forward-backward algorithm described later.

Finally, the processing device 11 determines the state q _t ^* (state at time t) included in the maximum likelihood state sequence Q ^* and measures the body temperature detection data O _t included in the body temperature data sequence O. status of the subject H0 to identify whether to correspond to any of the low temperature phase q ₁ and the high temperature phase q _2. Thereby, the processing apparatus 11 can estimate the state of the person to be measured H0 from the first day to the end date of the measurement date.

  And the processing apparatus 11 is provided with the homepage which can display the estimation result of the state of the to-be-measured person H0. For this reason, when the person to be measured H0 accesses the home page of the processing apparatus 11 using, for example, the mobile phone PT, the state estimation result is displayed on the mobile phone PT or the computer in response to the request of the person to be measured H0. Thereby, on the screen of the cellular phone PT or the like, for example, the estimation result of the state of the person being measured H0 over 6 months is displayed.

  The measured person H0 is not limited to the cellular phone PT, and may be configured to access the processing device 11 using various portable terminals, computers, or the like. Moreover, the processing apparatus 11 is good also as a structure which transmits the estimation result of a state toward the mail address registered previously.

  The menstrual cycle estimation apparatus according to the present embodiment has the above-described configuration. Next, the menstrual cycle estimation processing will be described with reference to FIG.

  Here, when the person to be measured H0 transfers the daily body temperature detection data D1 from the sensor device 1 to the processing device 11, the body temperature detection data D1 for a plurality of days (for 6 months) is stored in the processing device 11 in advance. It is assumed that And the processing apparatus 11 starts the below-mentioned menstrual cycle estimation process shown in FIG. 7 by receiving the request | requirement instruction | command etc. from the to-be-measured person H0.

  First, in step 1, the missing data is deleted from the accumulated body temperature detection data D1 for six months. At this time, the sensor device 1 reads the body temperature detection data D1 at intervals of 10 minutes over 6 hours between the start time ts and the end time te, and stores them in the storage unit 6. For this reason, the number of data of body temperature detection data D1 for one day is 37, and the number of data of body temperature detection data D1 for six months is about 6600. However, in reality, about 20 to 50% of missing data is actually generated. Therefore, the number of body temperature detection data D1 remaining after deletion of the missing data is about 3300 to 5200. For this reason, the length T of the body temperature data series O is also about 3300 to 5200 corresponding to the number of data after deletion of the missing data.

Next, in step 2, body temperature detection data D1 for the six months remaining in step 1 is connected to generate a body temperature data series O. At this time, the body temperature data series O is configured by connecting the body temperature detection data D1 on the start time ts side on the first measurement day to the body temperature detection data D1 on the end time te side on the measurement end date. Therefore, the body temperature detection data O _t (1 ≦ t ≦ T) included in the body temperature data series O is arranged in time series from the start time ts side of the first measurement day to the end time te side of the measurement end date.

  Next, in step 3, a filter using a median filter is applied to the body temperature data series O as a pre-process before performing this process consisting of a hidden Markov model construction process and a state series calculation process described later. Process. At this time, the median value filter substitutes, for example, 11 consecutive data values of these median values into time-series central data (sixth data arranged in time series). Thereby, an apparent outlier included in the body temperature data series O can be removed.

Next, in step 4, the body temperature detection data O _t is subjected to a discretization process. In this discretization process, for example, the body temperature detection data O _t included in the body temperature data series O is used by using a rounding-off method with a precision of 0.1 ° C., a vector quantization method, or a K-means method. Data clustering. Thereby, the body temperature detection data O _t becomes a numerical value expressed at intervals of 0.1 ° C. between 32 ° C. and 40 ° C., for example.

  Next, in step 5, a hidden Markov model construction process shown in FIG. 9 described later is performed, and a model HMM is constructed using the body temperature data series O after the pre-processing and the discretization process. Specifically, an optimal model parameter λopt (λopt = argmax [P (O | λ)]) of the model HMM is obtained so that the likelihood of the body temperature data series O becomes the highest.

Next, in step 6, for example, a state series calculation process using a forward-backward algorithm is performed. Specifically, based on the following equation (1), the model HMM constructed in step 5 calculates the state q _t ^* included in the maximum likelihood state sequence Q ^* that outputs the body temperature data sequence O. At this time, the state series Q ^* is a path of a state in which the model HMM is most likely to output the body temperature data series O. In Equation 1, γ _t (i) indicates the probability of being in state i at time t, and is calculated based on Equation 4 described later.

Next, in step 7, a state specifying process is performed to estimate the state of the person being measured H0 from the first day to the end date of the measurement date. Specifically, first, the filtering process using the median filter is performed again on the maximum likelihood state sequence Q ^* . At this time, for example, for 11 consecutive states, the median filter substitutes these median values into the middle (sixth from the front in time series order) state. That is, the median filter substitutes a state that occupies a majority of the low temperature phase q ₁ and the high temperature phase q ₂ included in the 11 consecutive states into the middle state.

The state q _t ^* of the subject H0 when the body temperature detection data O _t included in the body temperature data series O is measured from the maximum likelihood state series Q ^* output after the filter processing is the low temperature phase q ₁ and the high temperature phase. Specify which of q ₂ is applicable. Specifically, the number of the low-temperature phase q ₁ and the high-temperature phase q _{2 on} the same day is counted, and whether the day is in the low-temperature phase q ₁ or the high-temperature phase q ₂ is determined by a majority rule or other probability method. . Thereby, the state of the person to be measured H0 from the first day to the end date of the measurement date is estimated, and the two-phase body temperature in the menstrual cycle of the person to be measured H0 is specified.

  Finally, in step 8, the estimation result of the state of the person to be measured H0 from the first day to the end date of the measurement date specified in step 7 is displayed on the mobile phone PT and the computer of the person to be measured H0.

  Next, the hidden Markov model construction process will be described with reference to FIG. In the present embodiment, the case where the model parameter λ of the model HMM is calculated using the Balm-Welch learning algorithm will be described as an example.

In step 11, various coefficients necessary for constructing the model HMM are set. Specifically, the number N of hidden states included in the model HMM, the number M of possible measurement values (discretization number of the body temperature detection data O _t ), the length T of the body temperature data series O, and the body temperature data series O are respectively represented. Read.

At this time, since the number N of hidden states, for example, the model HMM shown in FIG. 7, includes the low temperature phase q ₁ and the high temperature phase q _{2, there} are two states. Further, the number M of possible measurement values is 81, for example, when 32 to 40 ° C. is expressed with an accuracy of 0.1 ° C. The length T of the body temperature data series O is about 3300 to 5200, for example, when the body temperature detection data O _t for 6 months after deleting the missing data.

Next, in step 12, an initial value is substituted for the model parameter λ. Here, as the initial value of the state transition probability a _ij, a value of about 1 / N is substituted on the assumption that the probability of transition to all states is substantially equal. However, if all the state transition probabilities a _ij are set to the same value, learning of the parameter λ is difficult to end (converge), so a slightly different value is substituted.

Further, the initial value of the body temperature detection data output probability b _j (k) has almost the same probability of outputting all possible body temperature discrete values in each state, that is, 32 to 40 ° C. as body temperature detection data O _{t in} each state. Assuming that the values are output uniformly, a value of about 1 / M is substituted. Further, as the initial value of the initial state probability π _i, a value of about 1 / N is substituted assuming that the probability of being in all the states is almost equal.

Next, in step 13, the forward variable α _t (i) is calculated based on the following equation (2). At this time, the forward variable α _t (i) is the probability that the state i is in the state i at time t after observing the partial observation sequence {O ₁ , O ₂ ,..., O _t } from time 1 to time t. Is shown.

Next, in step 14, the backward variable β _t (i) is calculated based on the following equation (3). At this time, the backward variable β _t (i) indicates the probability of observing the partial observation sequence {O _{t + 1} , O _{t + 2} ,..., O _T } from the time t + 1 to the last in the state i at the time t. ing.

Next, in step 15, the probability γ _t (i) of being in the state i at time t is calculated based on the following equation (4).

Next, in step 16, the parameter λ is updated using the probability γ _t (i). Specifically, it is as follows. First, assuming that the probability of being in state i at time t and being in state j at time t + 1 is ξ _t (i, j), this probability ξ _t (i, j) is expressed by the following _equation (5).

Here, from the definition of the forward variable α _t (i) and the backward variable β _t (i), between the probability γ _t (i) and the probability ξ _t (i, j), There is a relationship to show.

As described above, when the body temperature detection data D1 is discretized, the initial state probability π _i , the state transition probability a _ij and the body temperature detection data output probability b _j (k) after the re-estimation calculation are expressed by the following equation 7, It can be expressed by the formulas 8 and 9.

  In the Balm-Welch learning algorithm, the probability Pr (O | λ) of outputting the data series O by the model parameter λ after the re-estimation is output by re-estimating the data series O by the model parameter λ before the re-estimation. It is guaranteed that the probability is higher than Pr (O | λ). Therefore, the model parameter λ based on the body temperature data series O can be obtained by repeating the above steps 13 to 16.

  When step 16 ends, the process proceeds to step 17. In step 17, in order to determine whether or not the model parameter λ has converged, the logarithmic value η of the probability Pr (O | λ) based on the reestimated model parameter λ based on the following equation (10): Is calculated. The logarithmic operation is performed because the value of the probability Pr (O | λ) tends to be too small, so that it is easier to determine convergence by using the logarithm.

  Next, in step 18, it is determined whether the difference Δη between the logarithmic values η before and after the re-estimation is greater than a predetermined convergence determination value δ. If “YES” is determined in step 18, the difference between the probabilities Pr (O | λ) before and after re-estimation is large and the parameter λ has not converged, so the processing from step 13 is repeated.

  On the other hand, when “NO” is determined in step 18, it is considered that the difference between the probabilities Pr (O | λ) before and after the re-estimation is small and the parameter λ has converged. For this reason, the process proceeds to step 19 where the re-estimation calculation result at this time is output as the optimum model parameter λopt, and the process returns to step 20.

  In the present embodiment, the menstrual cycle estimation process and the hidden Markov model construction process described above are performed. Next, the estimation result of the state of the subject H0 when these processes are used will be described.

  Here, when the body temperature detection data D1 is measured over a measurement period of 6 months for a specific person H0, the body temperature H0 during the measurement period is estimated using the body temperature detection data D1 for this measurement period. This will be described as an example.

  First, the processing device 11 deletes the missing data from the body temperature detection data D1 for six months accumulated in advance and connects the remaining data D1. As a result, as shown by a characteristic line A1 in FIG. 10, a body temperature data series O in which about 5200 data D1 is connected is generated.

Next, after pre-processing and discretizing the body temperature data series O, a hidden Markov model is constructed. Thus, since the maximum likelihood model HMM that outputs the body temperature data series O is constructed, the maximum likelihood state series Q ^* that outputs the body temperature data series O is obtained using the constructed model HMM. As a result, as shown by the characteristic line A2 in FIG. 10, a state series Q ^* having a periodicity of about one month corresponding to the body temperature data series O can be obtained. However, in the state series Q ^* , an outlier occurs throughout the measurement period, and there is a portion where the low temperature phase q ₁ and the high temperature phase q ₂ are rapidly switched like a δ function.

Therefore, the processing device 11 performs a state specifying process using a median filter or other probability method on the state series Q ^* obtained using the model HMM. As a result, outliers can be removed. As shown by the characteristic line A3 in FIG. 10, the state series Q ^* has a periodicity in units of about one month between the low temperature phase q ₁ and the high temperature phase q ₂ . The two phases of the menstrual cycle can be clearly grasped.

Further, the changing point from the high temperature phase q ₂ to the low temperature phase q ₁ becomes a start signal of menstruation. Therefore, in the state series Q ^* estimated by the processing device 11 (characteristic line A3 in FIG. 10), the time when the high temperature phase q ₂ is switched to the low temperature phase q ₁ is compared with the time when menstruation actually occurs. . As a result, as shown in FIG. 10, the timing of switching from the high temperature phase q ₂ to the low temperature phase q ₁ in the state sequence Q ^* and the timing of the menstrual period are almost the same, and the menstrual cycle using the model HMM It was confirmed that the estimation of is effective.

Thus, according to the present embodiment, a model HMM having two states of the low temperature phase q ₁ and the high temperature phase q ₂ is constructed using the body temperature data series O, and the model HMM outputs the body temperature data series O. since determining the state sequence of the maximum likelihood Q ^*, by determining the state q _t ^* included in the state sequence Q ^*, to identify the status of the subject H0 across multiple days were measured body temperature data series O Can do. As a result, even if measurement artifacts occur in the body temperature data series O due to the influence of outside air or the like, it is possible to clearly grasp the biphasicity of the subject H0.

The processing device 11, because performs post-processing using the median filter or other probability approach to state sequence of the maximum likelihood Q ^*, removing the discontinuous off states included in the state sequence Q ^* Can do. As a result, the state does not change frequently in units of hours or days, and the biphasicity of the subject H0 that repeats the low temperature phase q ₁ and the high temperature phase q ₂ with a periodicity of about one month is clearly grasped. be able to.

  Further, in the present embodiment, the processing device 11 generates a body temperature data series O by deleting a value outside the allowable range at least one of the preset temperature and temperature change in the body temperature detection data D1. To do. For this reason, when constructing the model HMM, body temperature detection data D1 other than the allowable range of temperature or the allowable range of temperature change is not used. As a result, the abnormal value of the body temperature detection data D1 does not affect the model HMM, and when the state of the person to be measured H0 is estimated using the model HMM, the accuracy of the estimation result can be increased.

  Further, since the processing device 11 performs the filtering process using the median filter on the body temperature data series O before constructing the model HMM, it is possible to remove the discontinuous outliers included in the body temperature data series O. it can. Since the treatment apparatus 11 constructs the model HMM using the body temperature data series O after performing the filtering process, the model parameters of the model HMM are compared with the case where the body temperature data series O including an outlier is used. It is possible to prevent the outlier from affecting λ, and to improve the estimation accuracy of the state of the person being measured H0.

Further, since the processing device 11 constructs the model HMM in a state where the body temperature detection data O _t of the body temperature data series O has a discrete value with respect to the temperature, the model HMM corresponds to the discrete body temperature detection data O _t. It has a body temperature detection data output probability B consisting of a discrete probability distribution. Therefore, the maximum likelihood state series Q ^* can be obtained using the body temperature data series O composed of the body temperature detection data O _t discrete to the temperature and the discrete model HMM. Further, since the discrete hidden Markov model HMM is used, the time, storage capacity, and the like required for the hidden Markov model construction process can be reduced as compared with the case where the continuous hidden Markov model is used.

  Further, in the present embodiment, the casing 2 of the sensor device 1 is provided with the display unit 9 for transferring the body temperature detection data D1 stored in the storage unit 6 to the external processing device 11. For this reason, it is possible to transfer the body temperature detection data D1 in the storage unit 6 to the external processing device 11 using the two-dimensional code displayed on the display unit 9, and to construct a model HMM using the processing device 11. it can.

For this reason, for example, even when the model HMM is constructed using the body temperature detection data D1 for several months for the person H0 to be measured, the processing device 11 stores the large amount of body temperature detection data D1 in a large-capacity storage circuit. Since they can be stored and processed at high speed, the model HMM can be constructed and the maximum likelihood state sequence Q ^* can be calculated quickly.

  In the first embodiment, steps 1 to 4 in FIG. 8 are specific examples of data series generation means, steps 1 and 2 are specific examples of out-of-range data deletion means, and steps 3 and 4 are preprocessing means. Specific examples are shown respectively. Further, Step 5 in FIG. 8 and Steps 11 to 20 in FIG. 9 are specific examples of model construction means, Step 6 is a specific example of state series calculation means, and Step 7 is a specific example of state specifying means. .

  Further, in the preprocessing in step 3 and the state specifying process in step 7, the median filter is used. However, for example, any means that removes outliers (outliers) may be used. For example, a low-pass filter or the like is used. It is good also as a structure.

  Further, the median filter used for the pre-processing and the state specifying process is configured to perform the filtering process using, for example, 11 pieces of data. It is set accordingly.

  Next, FIG. 11 and FIG. 12 show a second embodiment of the present invention, and the feature of this embodiment is that the processing device has a body temperature detection data of a body temperature data series having a continuous value with respect to the temperature. There is a configuration that builds a hidden Markov model in the state. In the present embodiment, the same components as those in the first embodiment described above are denoted by the same reference numerals, and the description thereof is omitted.

21 is a processing device according to the present embodiment, and the processing device 21 comprises data series generation means, model construction means, state series calculation means, and state identification means in substantially the same manner as the processing apparatus 11 according to the first embodiment. To do. However, as shown in FIG. 12, the processing device 21 constructs a hidden Markov model HMM in a state where the body temperature detection data O _t (1 ≦ t ≦ T) of the body temperature data series O has a continuous value with respect to the temperature. This is different from the processing apparatus 11 according to the first embodiment.

Specifically, the processing device 21 does not perform data discretization processing. For this reason, the body temperature detection data O _t (1 ≦ t ≦ T) is a continuous value with respect to the temperature.

The model HMM has a body temperature detection data output probability B composed of a continuous probability distribution according to the continuous body temperature detection data O _t . Therefore, it is assumed that the body temperature detection data output probability B is based on the synthesis of M Gaussian distributions, and is expressed by the following equation (11). At this time, the limiting condition shown in the formula 12 is added.

When the model parameter λ is updated (reestimated) in accordance with the definition of the body temperature detection data output probability B, each parameter (mixing coefficient c _jm , mean vector μ _jm , covariance matrix U _jm ) of the Gaussian distribution density function G is It calculates | requires using the following formula | equation of 13-13.

Thus, in the second embodiment configured as described above, it is possible to obtain substantially the same operational effects as those of the first embodiment. In particular, in the second embodiment, the processing device 21 constructs a hidden Markov model HMM in a state where the body temperature detection data O _t of the body temperature data series O has a continuous value with respect to the temperature. Has a body temperature detection data output probability B composed of a continuous probability distribution according to the continuous body temperature detection data O _t . For this reason, the state series Q ^* can be obtained using the body temperature data series O composed of the body temperature detection data O _t continuous with the temperature and the continuous hidden Markov model HMM. Further, since the continuous hidden Markov model HMM is used, the accuracy of the state sequence Q ^* can be improved as compared with the case where the discrete hidden Markov model is used as in the first embodiment.

  Next, FIG. 13 and FIG. 14 show a third embodiment of the present invention. The feature of this embodiment is that the processing device displays the state of the person to be measured using a pie chart and a bar graph. It is to have done. In the present embodiment, the same components as those in the first embodiment described above are denoted by the same reference numerals, and the description thereof is omitted.

  FIG. 13 shows a pie chart 31 displayed on the homepage by the processing device 11. Here, the pie chart 31 is divided into first to eighth stages 31A to 31H having a sector shape, and half of the first to fourth stages 31A to 31D correspond to the low temperature period, and the remaining fifth to fifth stages 31A to 31D. The eighth stages 31E to 31H correspond to the high temperature period. Thereby, the pie chart 31 displays which stage the measured person corresponds to in the low temperature period or the high temperature period (first to eighth stages 31A to 31H).

  And one lap of the pie chart 31 corresponds to one cycle of the basal body temperature, and one cycle (one lap) is basically 28 days. At this time, the first, second, fifth, and sixth stages 31A, 31B, 31E, and 31F are allotted 4 days, respectively, and the third, fourth, seventh, and eighth stages 31C, 31D, 31G, and 31H are Each is assigned 3 days. Further, the first day of the first stage 31A is set, for example, as a menstrual start date recorded by the measured person H0 using the sensor device 1.

  And as a general rule, the processing device 11 selects any one of the first to fourth stages 31A to 31D depending on the number of days when the latest body temperature detection data D1 is measured (the latest measurement date) from the latest change point. Determine if it applies. That is, the processing apparatus 11 determines which of the first to fourth stages 31A to 31D corresponds to the number of days that the latest measurement date has elapsed from the first day of the first stage 31A in principle. Then, the processing device 11 displays, for example, a character symbol (for example, an asterisk) during the stage corresponding to the latest measurement date. Note that the corresponding stage is not limited to a character symbol, and may be displayed by blinking, for example.

However, after a lapse of a change point falling latest up, when the elapsed also change point from the low temperature phase q ₁ to the high-temperature phase q ₂ (rising change point) is the day when the rising change point has occurred fifth stage It is determined that the day has entered 31E. Then, it is determined which of the fifth to eighth stages 31E to 31H corresponds to the number of days since the latest measurement date has elapsed from the rising change point, and, for example, a character symbol is displayed in the corresponding stage.

  On the contrary, if the rising point does not occur even if the latest measurement date is 12 days or more after the first day of the first stage 31A, the character symbol is displayed in the fourth stage 31D, assuming that it remains in the fourth stage 31D. To do. In particular, when a rising change point does not occur even when the latest measurement date is 15 days or more after the first day of the first stage 31A, it is determined that the low temperature period has been prolonged, and this is displayed.

If the next falling change point does not occur even if 17 days or more have elapsed from the first day of the fifth stage 31E (date of the latest rising change point), the processing device 11 Character symbols are displayed during the stage 31H, and a message that there is a possibility of pregnancy is displayed. Here, when pregnant, the basal body temperature generally decreases gradually from about 16 weeks and returns to the low temperature phase q ₁ around 20 weeks, so the low temperature phase q ₁ between about 16 weeks and 20 weeks. When returning to, a message indicating that the possibility of pregnancy is high is displayed.

  FIG. 14 shows a bar graph 32 displayed on the homepage by the processing device 11 separately from the pie chart 31. Here, the bar graph 32 determines the day when the latest falling change point has occurred as the start date of menstruation, and displays it with a symbol (for example, a circle symbol) indicating the start of menstruation along with the date.

  In addition, the processing apparatus 11 displays one or a plurality of (for example, three) character symbols (for example, black star marks) indicating the high temperature period on days corresponding to the high temperature period. On the other hand, the processing device 11 displays one or more (for example, three) character symbols (for example, white star marks) indicating that the temperature is in the low temperature period on a day corresponding to the low temperature period.

  Thereby, compared with the case where the display by a line graph is used, it can be displayed on the screen of the cellular phone PT, for example, in an easy-to-see manner.

  The character symbol indicating the high temperature period may be displayed in different colors such as red, and the character symbol indicating the low temperature period may be blue.

  Further, the subject is notified of the physical condition (good, bad) of the day, and the bar graph 32 displays the physical condition of the day of the declaration at a position corresponding to the date with a simple character symbol (for example, good with φ, It is good also as a structure which displays it using (normal feeling with #, bad feeling with b).

  Moreover, it is good also as a structure which estimates the next menstrual period (falling change point) based on the past menstrual cycle, and displays this prediction date. In this case, the next menstrual period may be predicted by adding the number of days for one previous menstrual cycle to the previous menstrual period, and for the past several menstrual cycles, the average number of days for these one cycle is calculated. It may be predicted by calculating and adding this average number of days to the previous menstrual period. Furthermore, assuming that the menstrual cycle is subject to periodic fluctuations in the length of the menstrual cycle, the number of days for one cycle of the menstrual cycle is calculated in consideration of the periodicity of the menstrual cycle, and the number of days for this calculated cycle is the previous time. The next menstrual period may be predicted by adding to the menstrual period.

  Moreover, it is good also as a structure which displays that as the ovulation period in which the period of two days before and the day when the rising change point occurred is possible. Furthermore, it is good also as a structure which estimates the next ovulation period based on the past menstrual cycle. In this case, the next ovulation period can be predicted in substantially the same manner as the next menstrual period.

  In addition, when luteal dysfunction occurs in the measurement subject, since lutein hormone secretion is insufficient and the high temperature period cannot be maintained, body temperature falls in the latter half of the high temperature period, menstrual periods occur while the high temperature period is extremely short, It becomes difficult to get pregnant. For this reason, when the period of a high temperature period becomes less than 10 days, it is good also as a structure which estimates that there exists a possibility of a luteal dysfunction and displays that.

  In the anovulatory cycle, although there is periodic bleeding, changes in basal body temperature do not appear, and menstruation remains in the low temperature period. In the case of such a periodic pattern, there is no ovulation, so there is no pregnancy. For this reason, when the biphasic change of a menstrual cycle does not appear, it is good also as a structure which estimates that there exists a possibility of an anovulatory cycle and displays that.

  Thus, in the third embodiment configured as described above, it is possible to obtain substantially the same operational effects as those in the first embodiment. In particular, since the processing device 11 is configured to display the pie chart 31 and the bar graph 32 indicating the state of the person being measured, the state of the latest measurement date and the like when the person to be measured visually observes the pie chart 31 and the bar graph 32. Can be easily grasped.

  Moreover, in 3rd Embodiment, it was set as the structure which produces the pie chart 31 and the bar graph 32 using the processing apparatus 11 by 1st Embodiment. However, the present invention is not limited to this. For example, the pie chart 31 and the bar chart 32 may be created using the processing device 21 according to the second embodiment.

In each of the above embodiments, the hidden Markov model HMM is configured to have only two states of the low temperature phase q ₁ and the high temperature phase q ₂ . However, the present invention is not limited to this. For example, like the hidden Markov model HMM according to the modification shown in FIG. 15, the intermediate temperature phase q ₂ is provided between the low temperature phase q ₁ and the high temperature phase q _3, and the hidden Markov model HMM is provided. It is good also as a structure which has these three states. Thereby, the rising change point from the low temperature phase q ₁ to the high temperature phase q ₃ and the falling change point from the high temperature phase q ₃ to the low temperature phase q ₁ can be examined in detail. In addition, when the intermediate temperature phase q ₂ is provided, the probability of direct transition between the low temperature phase q ₁ and the high temperature phase q ₃ is considered to be low. Therefore, when obtaining the maximum likelihood state sequence, the Viterbi algorithm (Viterbi algorithm) algorithm) is effective in reducing calculation time and calculation cost.

  In each of the above embodiments, the sensor device 1 is configured to set the measurement start time and end time in advance and detect the body temperature detection data D1 at regular intervals between them. However, the present invention is not limited to this, and the sensor device 1 starts reading and storing the body temperature detection data D1 when the body temperature T1 detected by the body temperature detection unit 3 rises above a certain temperature (for example, 32 ° C.), for example. When the temperature falls below a certain temperature, the reading of the body temperature detection data D1 and the like may be terminated.

  Further, the sensor device 1 is provided with a micro azimuth switch, and when the person to be measured changes the posture from the sitting position or the standing position to the lying position, the azimuth switch is turned on, and the person to be measured changes the posture from the lying position to the sitting position or standing position. If changed, the azimuth switch may be turned off. As a result, the sensor device 1 can start and end reading of the body temperature detection data D1 in accordance with the on / off state of the azimuth switch.

  In each of the above embodiments, the body temperature detection data D1 during sleep is detected a plurality of times per day. However, for example, the body temperature detection data D1 may be detected only once per day after the body temperature is sufficiently stabilized.

It is an external view which shows the state which the to-be-measured person mounted | worn with the wearable sensor apparatus by the 1st Embodiment of this invention. It is a front view which shows the wearable sensor apparatus in FIG. It is a rear view which shows the wearable sensor apparatus in FIG. It is sectional drawing which looked at the wearable sensor apparatus from the arrow IV-IV direction in FIG. It is a block diagram which shows the structure of the control unit provided in the wearable sensor apparatus. It is explanatory drawing which shows the menstrual cycle estimation apparatus comprised from the wearable sensor apparatus and the external processing apparatus. It is explanatory drawing which shows the hidden Markov model by 1st Embodiment. It is a flowchart which shows the menstrual cycle estimation process. It is a flowchart which shows the construction process of the hidden Markov model in FIG. It is a characteristic diagram which shows the body temperature data series after pre-processing, the maximum likelihood state series, and the state series after state specific processing, respectively. It is explanatory drawing which shows the menstrual cycle estimation apparatus by 2nd Embodiment. It is explanatory drawing which shows the hidden Markov model by 2nd Embodiment. It is explanatory drawing which shows the pie chart by 3rd Embodiment. It is explanatory drawing which shows the bar graph by 3rd Embodiment. It is explanatory drawing which shows the hidden Markov model by a modification.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Wearable sensor apparatus 2 Casing 3 Body temperature detection part (body temperature detection means)
5 Control unit (data reading means)
6 Storage unit (storage means)
9 Display (transfer means)
11, 21 processing equipment

Claims (9)

A menstrual cycle estimation device that estimates whether a subject is in a high temperature period or a low temperature period using a body temperature data series obtained by measuring the body temperature of the subject over a plurality of days,
A hidden Markov model having two states, a high temperature phase and a low temperature phase, is constructed using the body temperature data series, the hidden Markov model obtains a maximum likelihood state series that outputs the body temperature data series, and the state series is A menstrual cycle estimation device configured to be used to identify the biphasic body temperature in the menstrual cycle of the measurement subject.   A casing that can be always attached to the body of the measurement subject, body temperature detection means that is provided on the body surface side of the measurement subject and detects the body temperature, and between the start time and the end time for each measurement day Reading means for reading body temperature detection data by the body temperature detection means at predetermined time intervals, storage means for storing body temperature detection data by the reading means, and body temperature detection data stored in the storage means for a plurality of days A data series generating means for generating the body temperature data series by concatenating, a model construction means for constructing the hidden Markov model using a body temperature data series by the data series creation means, and the hidden Markov model by the model construction means A state sequence calculating means for obtaining a maximum likelihood state sequence for the model to output the body temperature data series, and a state sequence obtained by the state sequence calculating means. Menstrual cycle estimator of claim 1 comprising a structure and a state specifying means for specifying a biphasic temperature in menstrual cycle makers.   The state specifying means performs a filtering process using a median filter on the state series, and the state included in the state series after the filtering process is the state of the subject when each body temperature detection data is measured. The menstrual cycle estimation apparatus according to claim 2, wherein the menstrual cycle estimation apparatus is configured as described above. The data series generation means deletes values outside the preset temperature and allowable temperature change range from the body temperature detection data, and generates a body temperature data series by connecting the remaining body temperature detection data. Means, and preprocessing means for performing filter processing using a median filter on the body temperature data series generated using the out-of-range data deletion means,
The menstrual cycle estimation device according to claim 2 or 3, wherein the model construction unit constructs the hidden Markov model using a body temperature data series after the filter processing by the preprocessing unit is performed.   5. The menstrual cycle estimation according to claim 2, 3, or 4, wherein the model constructing unit is configured to construct the hidden Markov model in a state where body temperature detection data of the body temperature data series has discrete values with respect to temperature. apparatus.   5. The menstrual cycle estimation according to claim 2, 3, or 4, wherein the model constructing unit is configured to construct the hidden Markov model in a state in which body temperature detection data of the body temperature data series has a continuous value with respect to temperature. apparatus. The model construction means is constituted by an external processing device,
The menstrual cycle estimation apparatus according to claim 2, 3, 4, 5 or 6, wherein the casing is provided with a transfer means for transferring body temperature detection data stored in the storage means to the external processing device. It is a menstrual cycle estimation method for estimating whether a measurement person is in a high temperature period or a low temperature period using a body temperature data series obtained by measuring the temperature of the measurement person over a plurality of days,
A hidden Markov model having two states, a high temperature phase and a low temperature phase, is constructed using the body temperature data series, the hidden Markov model obtains a maximum likelihood state series that outputs the body temperature data series, and the state series is A menstrual cycle estimation method that is used to identify the body temperature biphasic in the menstrual cycle of the subject. It is a menstrual cycle estimation method for estimating which state of the polymorphism of body temperature in the menstrual cycle using the body temperature data series obtained by measuring the body temperature of the subject over a plurality of days,
Using the body temperature data series to construct a hidden Markov model having a plurality of hidden states corresponding to body temperature polymorphism, the hidden Markov model obtains a maximum likelihood state series that outputs the body temperature data series, the state A menstrual cycle estimation method configured to identify polymorphisms of body temperature in a menstrual cycle of a measurement subject using a series.

Images