WO2025255587A1

WO2025255587A1 - Systems and methods for generating individual health and wellness insights

Info

Publication number: WO2025255587A1
Application number: PCT/US2025/032936
Authority: WO
Inventors: Andrew Heymann; Nora TOPHOF; Kash KAPADIA; Tali Treibitz; Rafael Rexach; Vaibhav AYACHIT
Original assignee: Kohler Co
Current assignee: Kohler Co
Priority date: 2024-06-07
Filing date: 2025-06-09
Publication date: 2025-12-11
Anticipated expiration: 2026-12-07

Abstract

The present disclosure relates to a system for evaluating health and wellness trends in an individual. The system includes one or more input devices to collect at least one health input from an individual and transmit the at least one health input to a computing device. The computing device receives the at least one health input from the individual. The computing device includes a processor and a device comprising a non-transitory storage medium encoded with instructions executable by a processor that cause the processor to: store the at least one health input in a memory; identify one or more relationships between the at least one health input and the general population data; determine when the one or more relationships indicate the individual may experience a health event; and generate an alert when the one or more relationships indicate the individual may experience a health event.

Description

SYSTEMS AND METHODS FOR GENERATING INDIVIDUAL HEALTH AND WELLNESS INSIGHTS

TECHNICAL FIELD

[0001] The present disclosure relates generally to generating individual health and wellness insights. More specifically, this application relates to collecting health input data from one or more input devices or streams, comparing the health input data to general population data or an individual’s historic baseline data to identify one or more relationships or insights, and/or evaluate health and wellness trends based on the one or more identified relationships.

BRIEF DESCRIPTION OF THE DRAWINGS

[0002] The disclosure will become more fully understood from the following detailed description, taken in conjunction with the accompanying figure, wherein like reference numerals refer to like elements, in which:

[0003] FIG. l is a schematic illustration of an example apparatus for analyzing a bodily emission, according to at least one embodiment of the present disclosure.

[0004] FIG. 2 is a schematic diagram of an example system for evaluating health and wellness trends in an individual, according to at least one embodiment of the present disclosure;

[0005] FIG. 3 is a method flow diagram illustrating a method for evaluating health and wellness trends in an individual, according to at least one embodiment of the present disclosure;

[0006] FIG. 4 is a schematic diagram of an example system of sensors for collecting at least one health input, according to at least one embodiment of the present disclosure;

[0007] FIG. 5 is a perspective view of a toilet seat , according to at least one embodiment of the present disclosure,

[0008] FIG. 6 is a method flow diagram of a method for training a machine learning model, according to at least one embodiment of the present di closure;

[0009] FIG. 7 is a method flow diagram of a method for generating individual health and wellness insights, according to at least one embodiment of the present disclosure;

[0010] FIG. 8 is a method flow diagram of a method for a user to use a healthy and wellness system, according to at least one embodiment of the present disclosure;

[0011] FIG. 9 is a schematic view of a user interface, according to at least one embodiment of the present disclosure; [0012] FIG. 10 is another embodiment of a schematic view of a user interface, according to at least one embodiment of the present disclosure.

DETAILED DESCRIPTION OF THE DRAWINGS

[0013] Collecting health data from sensor inputs, such as heart rate, blood pressure, and oxygen saturation, has been enabled by advances in wearable technology and mobile applications. Many patients use devices that allow for continuous monitoring and easy access to their vital signs over time. This ongoing data collection provides more opportunities for utilizing such data to help users understand health and wellness.

[0014] In a first aspect of the present disclosure, a system comprises one or more input devices for collecting at least one health input from the individual and transmit the at least one health input to a computing device via a network; the computing device for receiving the at least one health input from the individual, the computing device comprising: a processor; and a device comprising a non-transitory storage medium encoded with instructions executable by the processor which, when executed by the processor, cause the processor to: store the at least one health input in a device memory, wherein the device memory comprises general population data from a population of individuals and historical health data for an individual; identify one or more relationships between the at least one health input and the general population data or the historical health data; determine when the one or more relationships indicate the individual may experience a health event; and generate an alert when the one or more relationships indicate the individual may experience a health event. [0015] In a second aspect of the first aspect or any other aspect of the present disclosure, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: transmit the alert to a mobile device assigned to the individual.

[0016] In a third aspect of the first aspect or any other aspect of the present disclosure, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: request a manual input from the individual relating to one or more factors that affect the at least one health input of the individual.

[0017] In a fourth aspect of the third aspect or any other aspect of the present disclosure, wherein the one or more factors comprise at least one of a weather report, an exercise log, a mood log, a sleep log, and a nutrition log. [0018] In a fifth aspect of the first aspect or any other aspect of the present disclosure, wherein the one or more input devices comprise at least at least one sensor.

[0019] In a sixth aspect of the fifth aspect or any other aspect of the present disclosure, wherein the at least one sensor comprises at least one of a temperature sensor, a weight sensor, a pressure sensor, a urine analysis sensor, and an environment sensor, and a biometric sensor a PPG sensor, an ECG sensor, a camera, a microphone, a radar device, a smell sensor, or a stool sensor.

[0020] In a seventh aspect of the first aspect or any other aspect of the present disclosure, wherein the computing device receives at least one manual input from a user.

[0021] In an eighth aspect of the seventh aspect or any other aspect of the present disclosure, wherein the manual input includes information captured by the mobile device assigned to the user.

[0022] In a ninth aspect of the first aspect or any other aspect of the present disclosure, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: perform one or more pre-processing operations to curate the at least one health input, wherein the one or more pre-processing operations include at least one of synchronizing the at least one health input based on timestamp alignment, performing one or more data cleaning operations, or performing one or more modality specific feature extractions.

[0023] In a tenth aspect of the first aspect or any other aspect of the present disclosure, wherein: the one or more relationships between the at least one health input and the general population data or the historical health data are generated using one or more statistical analysis operations; and the one or more statistical analysis operations include at least one of a univariate statistical measurement, a bivariate relationship technique, a multivariate statistical technique, or an advanced statistical model.

[0024] In an eleventh aspect of the tenth aspect or any other aspect of the present disclosure, wherein the bivariate correlation technique includes at least one of a Pearson correlation coefficient, a Spearman's rank correlation, a Kendall's Tau, or scatter plots.

[0025] In a twelfth aspect of the first aspect or any other aspect of the present disclosure, wherein the multivariate statistical technique includes at least one of: a correlation matrix or heatmap, a Principal Component Analysis, a Factor Analysis, a Canonical Correlation Analysis, a regression analysis including multiple linear regression, or a time series specific statistical methods including a cross-correlation function and an autoregressive integrated moving average.

[0026] In a thirteenth aspect of the present disclosure, a method comprises receiving at least one health input from one or more input devices via a network; storing the at least one health input in a device memory, wherein the device memory comprises a general population data from a population of individuals; identifying one or more relationships between the at least one health input and the general population data or the historical health data from a user; determining when the one or more relationships indicate the individual may experience a health event; and generating an alert when the one or more relationships indicate the individual may experience a health event.

[0027] In a fourteenth aspect of the thirteenth aspect or any other aspect of the present disclosure, further comprising transmitting the alert to a mobile device assigned to the individual.

[0028] In a fifteenth aspect of the thirteenth aspect or any other aspect of the present disclosure, further comprising receiving at least one manual input from a user.

[0029] In a sixteenth aspect of the thirteenth aspect or any other aspect of the present disclosure, further comprising generating the one or more relationships between the at least one health input and the general population data or the historical health data using one or more statistical analysis operations, wherein the one or more statistical analysis operations include at least one of a univariate statistical measurement, a bivariate correlation technique, a multivariate statistical technique, or an advanced statistical model.

[0030] In a seventeenth aspect of the thirteenth aspect or any other aspect of the present disclosure, wherein the bivariate correlation technique includes at least one of a Pearson correlation coefficient, a Spearman's rank correlation, a Kendall's Tau, or scatter plots. [0031] In an eighteenth aspect of the sixteenth aspect or any other aspect of the present disclosure, wherein the multivariate statistical technique includes at least one of: a correlation matrix or heatmap, a Principal Component Analysis, a Factor Analysis, a Canonical Correlation Analysis, a regression analysis including multiple linear regression, or a time series specific statistical methods including a cross-correlation function and an autoregressive integrated moving average.

[0032] In a nineteenth aspect of the present disclosure, a method for training a machine learning model to generate one or more health and wellness insights, comprising: providing health input data comprising general population data or the historical health data; performing one or more pre-processing steps to curate the health input data identifying one or more relationships between the at least one health input and the general population data or the historical health data from a user; determining when the one or more relationships indicate the individual may experience a health event; determining whether the one or more relationships exceed a predetermined threshold value; and locking the machine learning model.

[0033] In a twentieth aspect of the nineteenth aspect or any other aspect of the present disclosure, further comprising providing one or more labels to train the machine learning model using supervised learning.

[0034] The present disclosure relates to a system for evaluating health and wellness trends in an individual. The system includes one or more input devices that collect at least one health input from an individual and transmit the at least one health input to a computing device. The computing device receives the at least one health input from the individual. The computing device includes a processor and a device comprising a non-transitory storage medium encoded with instructions executable by a processor that cause the processor to: store the at least one health input in a device memory; identify one or more relationships between the at least one health input and the general population data; determine when the one or more relationships indicate the individual may experience a health event; and generate an alert when the one or more relationships indicate the individual may experience a health event. [0035] Furthermore, the present disclosure relates to a method for evaluating health and wellness trends in an individual, the method includes: receiving at least one health input from one or more input devices via a network; storing the at least one health input in a device memory, wherein the device memory comprises a general population data from a population of individuals; identifying one or more relationships between the at least one health input and the general population data; determining when the one or more relationships indicate the individual may experience a health event; and generating an alert when the one or more relationships indicate the individual may experience a health event.

[0036] Before turning to the figures, which illustrate the exemplary embodiments in detail, it should be understood that the present disclosure is not limited to the details or methodology set forth in the description or illustrated in the figures. It should also be understood that the terminology used herein is for the purpose of description only and should not be regarded as limiting.

[0037] Reference is now made to FIG. 1, which is a schematic illustration of apparatus 20 for analyzing a bodily emission, in accordance with some applications of the present invention. As shown, apparatus 20 typically includes a sensor module 22, which is placed inside a toilet bowl 23. For some applications (not shown), the sensor module (and/or additional components of the apparatus) is integrated into the toilet bowl. The sensor module includes an imaging component 24, which in turn includes one or more light sensors that are configured to receive light from bodily emissions (typically, urine or feces 26) that were emitted by the subject and are disposed inside the toilet bowl. For example, the light sensors may include a spectrometer, or may include one or more cameras, as described in further detail hereinbelow. A computer processor analyzes the received light and determines whether there is a presence of blood inside the toilet bowl. Typically, the computer processor detects one or more spectral components within the received light that are indicative of light absorption by a component of erythrocytes, by analyzing the received light (e.g., by performing spectral analysis on the received light). (Such spectral components are referred to herein as examples of a blood signature, since certain combinations of such components, as described herein, are indicative of the presence of blood.) Further typically, the steps of receiving light, analyzing the received light, and determining whether there is a presence of blood inside the toilet bowl are performed without requiring any action to be performed by any person (e.g., the user, a caregiver, or a healthcare professional) subsequent to the subject emitting the bodily emission into the toilet bowl.

[0038] For some applications, apparatus 20 includes a power source 28 (e.g., a battery pack), that is disposed outside the toilet bowl inside a housing 30, as shown in FIG. 1 . Alternatively or additionally, the sensor module is connected to mains electricity (not shown). Typically, the power source and sensor module 22 are connected wiredly (as shown), or wirelessly (not shown). In accordance with respective applications, the computer processor that performs the above described analysis is disposed inside the toilet bowel (e.g., inside the same housing as the sensor module), inside housing 30, or remotely. For example, as shown, the sensor module may communicate wirelessly with a user interface device 32 that includes a computer processor. Such a user interface device may include, but is not limited to, a phone 34, a tablet computer 36, a laptop computer 38, or a different sort of personal computing device. The user interface device typically acts as both an input device and an output device, via which the user interacts with sensor module 22. The sensor module may transmit data to the user interface device and the user interface device computer processor may run a program that is configured to analyze the light received by the imaging module and to thereby detect whether there is a presence of blood inside the toilet bowl.

[0039] For some applications, sensor module 22 and/or the user interface device communicates with a remote server. For example, the apparatus may communicate with a physician or an insurance company over a communication network without intervention from the patient. The physician or the insurance company may evaluate the results and determine whether further testing or intervention is appropriate for the patient. For some applications, data relating to the received light are stored in a memory (such as memory 46 described hereinbelow). For example, the memory may be disposed inside the toilet bowel (e.g., inside the sensor unit), inside housing 30, or remotely. Periodically, the subject may submit the stored data to a facility, such as a healthcare facility (e.g., a physician's office, or a pharmacy) or an insurance company, and a computer processor at the facility may then perform the above-described analysis on a batch of data relating to a plurality of bodily emissions of the subject that were acquired over a period of time.

[0040] It is noted that the apparatus and methods described herein include a test in which the subject is not required to physically touch the bodily emission. Furthermore, the subject is typically only required to touch any portion of the dedicated sensing apparatus periodically, for example, in order to install the device, or to change or recharge the device batteries. (It is noted that the subject may handle the user interface device, but this is typically a device (such as a phone) that subject handles even when not using the sensing apparatus.) In certain embodiments, the apparatus and methods described herein do not require adding anything to the toilet bowl subsequent to the subject emitting a bodily emission into the toilet bowl, in order to facilitate the spectral analysis of the emission, and/or a determination that the bowl contains blood. For some applications, the subject is not required to perform any action after installation of the apparatus in the toilet bowl. The testing is automatic and handled by the apparatus, and monitoring of the subject's emissions is seamless to the subject and does not require compliance by the subject.

[0041] In certain embodiments, subsequent to the subject emitting a bodily emission into the toilet bowl (and typically once the subject has finished excreting the bodily emission, and the bodily emission is at least partially disposed within the water of the toilet bowl), the bodily emission is imaged by receiving reflected and/or transmitted light from the toilet bowl, without requiring any action to be performed by any person subsequent to the emission. It is noted that for some applications, an input is requested from the subject, via the user interface device, if an indication of the presence of blood in the toilet bowl is detected.

[0042] For some applications, for each emission of the subject, in the case of positive signal, the apparatus reports the finding to the patient via an output device, e.g., via user interface device 32. For some applications, the output device includes an output component (such as a light (e.g., an LED) or a screen) that is built into apparatus 20. For some applications, the data are analyzed locally but the results are transmitted to the healthcare provider or to insurance carrier over a network connection.

[0043] For some applications, the apparatus monitors bodily emissions of the subject over an extended period of time, e.g., over more than one week, or more than one month. In certain embodiments, the system described herein analyzes input data from one or more sensors to determine whether blood is present in the bowl.

[0044] In certain embodiments, the apparatus monitors other health inputs, such as hydration and gut health. For example, sensors integrated within the system, such as urine analysis sensors, stool sensors, and volatile organic compound (VOC) sensors, may automatically assess hydration status by measuring parameters like urine specific gravity, color, and volume. Similarly, gut health can be evaluated through the analysis of stool characteristics, including color, consistency, and the presence of blood or other biomarkers, as well as the detection of specific gases or compounds associated with digestive health using E-nose or smell sensors. The resulting data from these sensors can be used to identify trends, deviations from individual baselines, or early signs of gastrointestinal or hydration-related health issues, enabling the system to generate personalized health insights and recommendations for the user. Sensor systems used herein are illustrated and described in further detail with respect to FIGS. 2 and 4.

[0045] For some applications, the computer processor which analyzes the received light utilizes machine learning techniques, such as anomaly detection and/or outlier detection. For example, the computer processor may be configured to perform individualized anomaly detection or outlier detection that learns the patterns of output signals from each subject and detects changes in the characteristic blood signature of the subject. As described hereinabove, for some applications, the computer processor that performs the analysis is remote from and/or separate from the sensor module. For some applications, the sensor module is disposable, but even after disposal of the sensor module the computer processor has access to historic data relating to the subject, such that the historic data can be utilized in the machine learning techniques.

[0046] FIG. 2 is a schematic diagram of an example system 100 for evaluating health and wellness trends in an individual.

[0047] As used herein, a health event refers broadly to any occurrence or incident that affects an individual's health status that is deemed to be significant to an individual because it differentiates from relevant health input data. Relevant health input data may include an individual’s historic baseline health data or benchmark data from a larger number of comparable individuals. Various factors that are measurable by one or more input devices 102 may contribute to determining whether an individual is experiencing a health event. It is desirable to measure these various factors using the one or more input devices 102, compare the various factors to a much larger data set to identify trends, and generate one or more alerts that notify an individual that they are likely experiencing a health event.

[0048] Furthermore, systems and methods described herein for generating health and wellness insights include analyzing trends within input data from one or more sensors to draw comparisons between an individual’s baseline health or benchmark data.

[0049] The system 100 includes one or more input devices 102 for collecting data and transmitting the data over a network 120.

[0050] In certain embodiments, network transmission can be performed through a variety of methods, such as wired connections (including Ethernet cables and fiber optics), wireless technologies (such as Wi-Fi, Bluetooth, and cellular networks), satellite communication, infrared transmission, and radio frequency (RF) links. Furthermore, in certain embodiments, the network 120 may include transmissions between one or more sensors 106 and an internal processor, such as the device processor 105, the mobile device processor 111, or the computing device processor 142.

[0051] In certain embodiments, the one or more input devices 102 include at least one device 104. The Device 104 collects at least one health input 109 from the individual, using one or more sensors 106, and stores the at least one health input 109 in a Device memory 108.

[0052] Furthermore, in certain embodiments, the system 100 includes the apparatus 20 which may communicate wiredly or wirelessly with the computing device 140. In certain embodiments, the input devices 102 include the apparatus 20 and one or more sensors associated with the apparatus to collect health input data 109 and store the health input data 109 in a memory, such as the device memory.

[0053] Data and other functions described herein, including the health input data 109 and general population data 134, may be stored as one or more instructions 146 or code on a computer-readable medium (including, by non-limiting example, the Device memory 108, the mobile device memory 112, and the computing device memory 144) and executed by a hardware-based processing unit (including, by non-limiting example, the KD device processor 105, the mobile device processor 111, and the computing device processor 142). Computer-readable media may include non-transitory computer-readable media, which corresponds to a tangible medium such as data storage media (e.g., RAM, ROM, EEPROM, flash memory, or any other medium that can be used to store desired program code in the form of instructions or data structures and that can be accessed by a computer).

[0054] Instructions may be executed by one or more processors, such as one or more digital signal processors (DSPs), general purpose microprocessors, application specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuitry. Accordingly, the term “processor” as used herein may refer to any of the foregoing structure or any other physical structure suitable for implementation of the described techniques. Also, the techniques could be fully implemented in one or more circuits or logic elements.

[0055] As used herein, health input data 109 (or health input(s)) broadly refers to any information that can be used to assess the health of an individual. This may include, by nonlimiting example, information relating to an individual’s vital signs of an individual, physical activity, sleep patterns, nutritional intake, patient health history, self-reported symptoms, biometric measurements (e.g., height, weight, body mass index, measurements of various body parts, body fat percentage, etc.), blood glucose measurements, environmental exposure, and mental health metrics such as stress levels, mood fluctuations, and symptoms of anxiety or depression.

[0056] As described herein, the device memory 108 stores health input data 109 collected by the one or more sensors 106. The health input data 109 may be transmitted over a network 120 to a computing device 140 by the device processor 105.

[0057] In certain embodiments, the one or more input devices 102 include a mobile device 110. The mobile device 110 can include a mobile device processor 111 and a mobile device memory 112. The mobile device processor 111 is executes instructions stored in the mobile device memory 112 to perform one or more tasks relating to the collection of the health input data 109.

[0058] In certain examples, the mobile device 110 automatically captures health input data 109. The automatic capture of health input data 109 by the mobile device 110 may include, by non-limiting example, obtaining information from peripheral devices connected to the mobile device 110 via the network 120 (e.g., a vital sign monitor), obtaining weather metrics from weather data providers, and obtaining protected health information (PHI) from the individual’s health records.

[0059] In certain examples, the mobile device 110 receives at least one manual input from a user to receive health input data 109. This may include, by non-limiting example, information relating to one or more factors that affect at least one health input 109 of the individual, such as a personal weather report, an exercise log, a mood log, a sleep log, and a nutrition log. The personal weather report may include information manually inputted by the individual relating to the weather for a predefined period of time (e.g., weather conditions, time spent outside, etc.). The exercise log may include information relating to the quantity of times the individual exercises during a predefined period of time, the time spent during each exercise session, and the activities the individual performs during each exercise session. The mood log may include information pertaining to the individual’s mood and overall mental health, including symptoms relating to stress, anxiety, depression, or other mental health disorders. The sleep log may include information relating to the individual’s time spent sleeping each night, a perceived quality of sleep, and symptoms the individual experiences relating to the amount of sleep the individual experiences during a predefined period of time. The nutrition log may include a description of food and beverages consumed during a predefined period of time, the timing of each meal, an overall caloric intake during a predefined period of time, a macronutrient breakdown of the food that is consumed during a predefined period of time, water intake during a predefined period of time, symptoms of digestive distress and descriptions of overall gut health, symptoms relating to the amount of food consumed and any adverse reactions, etc. These manual entries can be entered into the mobile device memory 112 through one or more applications as part of a diary function or self-reporting function.

[0060] In certain embodiments, the nutrition log can be updated using the camera of the mobile device 110 by capturing an image of food that the individual is consuming. Further, the camera of the mobile device 110 may be used to capture information from menus that describe the food the individual is consuming. Information from a menu may be automatically captured through the use of one or more methods including a QR scanner, a barcode scanner, a software development kit, or a third-party application.

[0061] Furthermore, in certain embodiments, the health input data 109 may be stored in the form of photographs taken from a camera integrated with the mobile device 110 or separate from the mobile device 110.

[0062] The system 100 can include general population data 134 within a larger dataset of third-party data 130. In certain embodiments, the third-party data 130 includes a transmission device for transmitting the general population data 134 to a computing device 140 to be used to compare with health input data 109 received from the individual. A process by which health input data 109 is compared to general population data 134 to determine various relationships is described herein with respect to FIGS. 1-3. [0063] The system 100 includes a computing device 140 for receiving and analyzing the health input data 109 and general population data 134. The health input data 109 and the general population data 134 are received at a network interface 150 and are stored within the computing device memory 144.

[0064] The computing device memory 144 includes a non-transitory storage medium encoded with instructions 146 executable by the computing device processor 142 which, when executed by the computing device processor 142, cause the computing device processor 142 to perform various functions as described below with respect to FIG. 3. In certain embodiments, one or more of the functions described herein are performed using an artificial intelligence component 148, which is also described below with respect to FIG. 3.

[0065] FIG. 3 is a method flow diagram illustrating a method 200 for evaluating health and wellness trends in an individual and storing the at least one health input 109 in a device memory 108.

[0066] The method 200 includes a step 210 of receiving at least one health input 109 from the one or more input devices 102. As described above with respect to FIG. 2, the at least one health input 109 is collected by the one or more input devices 102 and transmitted to the computing device 140 over a network 120. In certain embodiments, step 210 may include receiving general population data 134 from third-party data 130. In certain embodiments, step 210 includes receiving health input data 109 from a plurality of individuals over time that generates general population data 134 without requiring the general population data 134 to be received via the third-party data 130.

[0067] The method 200 includes a step 220 of storing the at least one health input 109. The at least one health input 109 is received by the computing device 140 at the network interface 150 and stored within the computing device memory 144 by the computing device processor 142. In certain embodiments, step 220 includes storing the general population data 134 within the computing device memory 144.

[0068] The method 200 includes a step 230 of identifying one or more relationships between the at least one health input 109 and the general population data 134. The one or more relationships between the at least one health input 109 and the general population data 134 can be identified using a variety of methods, including, by non-limiting example, aggregation of collected data from the one or more input devices 102, data benchmarking of the health input data 109 with the general population data 134 to identify differences, similarities, and deviations from a general population, statistical analysis that analyzes a relationship between the health input data 109 and the general population data 134 (e.g., the use of regression analysis, a calculation of relationship coefficients, or the testing of various hypotheses within the data set), an identification of trends or deviations from a norm for one or more forms of health input data 109, and an interpretation of an individual’s overall health and wellness trends through the extraction of health insights that can allow a healthcare professional to make health recommendations.

[0069] In certain embodiments, step 230 can also include identifying one or more relationships between the at least one input 109, the general population data 134, and a stored memory of an individual’s health inputs over time (a.k.a. historic health data). This may include longitudinal analysis of the individual’s health input data 109 to establish a personalized baseline and to detect significant deviations or trends over time. Methods may include calculating the difference between current and baseline measurements, identifying statistically significant changes, and monitoring for outlier events relative to the individual’s historic patterns. Such comparisons enable the system to flag new or emerging health risks unique to the individual, provide health and wellness trends, and generate tailored insights or recommendations that account for both the person’s prior health status and broader population trends.

[0070] In certain embodiments, the identification of one or more relationships is performed using an artificial intelligence component 148. The artificial intelligence component 148 identifies the one or more relationships and extract insights from the one or more relationships based on the health input data 109 and the general population data 134. In certain embodiments, the artificial intelligence component 148 is trained from a labeled dataset using supervised learning to classify data using various methods, including, by nonlimiting example, linear regression, logistic regression, and neural networks. In other examples, the artificial intelligence component 148 can be trained using unsupervised learning to identify the relationships between the health input data 109 and the general population data 134 without the use of a labeled dataset.

[0071] In certain embodiments, supervised learning may be used by the artificial intelligence component 148 to train a machine learning model. As described above, supervised learning approaches involve training algorithms on labeled datasets, where the desired output for each input is known, enabling the model to learn mappings between health input data and specific outcomes or risk categories.

[0072] A variety of supervised learning algorithms may be employed, including regression techniques for predicting continuous health metrics and classification algorithms such as support vector machines (SVM), random forests, and k-nearest neighbors (KNN) for assigning input data to discrete health states or risk levels. Hidden Markov Models (HMMs) may also be used for sequential or time-dependent health data, providing probabilistic modeling of temporal patterns.

[0073] In addition, neural network architectures can be utilized to capture complex, nonlinear relationships within the data. Multi-layer perceptrons (MLPs) are suitable for general- purpose regression and classification tasks, while recurrent neural networks (RNNs) and long short-term memory networks (LSTMs) are particularly effective for modeling sequential and time-series health data. Convolutional neural networks (CNNs) may be applied to sensor data with spatial or structured patterns, such as those arising from imaging or multi-dimensional sensor arrays. By leveraging these supervised learning techniques, the system can improve the accuracy and robustness of health risk predictions and personalized wellness insights.

[0074] In certain embodiments, unsupervised machine learning may be used by the artificial intelligence component 148 to train a machine learning model. Methods such as Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE) can be applied for dimensionality reduction and visualization, helping to uncover underlying patterns or clusters within the data. Additionally, clustering algorithms like K-means may be used to automatically group similar health profiles or detect novel patterns, enabling the system to identify subpopulations, trends, or anomalies that may not be apparent through supervised approaches.

[0075] In certain embodiments, foundation models that are pre-trained using multimodal data can be leveraged for developing one or more machine learning models.

[0076] In certain embodiments, one or more techniques can be used by the system 100 to perform statistical analysis. These statistical techniques may include, by non-limiting example, univariate statistical measurements, bivariate correlation techniques, multivariate statistical techniques, and advanced statistical modeling.

[0077] Univariate statistical measurements may be applied to each sensor input to summarize and describe the distribution of data collected from individual modalities. Common metrics such as mean, median, standard deviation, minimum and maximum values, and histograms are calculated to provide basic descriptive statistics for each health input. These measurements help establish baselines, detect outliers, and offer an initial overview of the user’s health data from each sensor.

[0078] Bivariate correlation techniques may be employed to explore pairwise relationships between two different health inputs or sensor modalities. Methods such as the Pearson correlation coefficient, Spearman’s rank correlation, and Kendall’s Tau are used to quantify the strength and direction of associations between variables. Visualizations like scatter plots may also be generated to aid in interpreting these relationships, supporting the identification of dependencies or patterns between pairs of health metrics.

[0079] Multivariate statistical techniques are utilized to analyze relationships among multiple sensors or health inputs simultaneously. These may include constructing correlation matrices or heatmaps to visualize interconnections, applying dimensionality reduction methods such as Principal Component Analysis (PCA) or Factor Analysis (FA), and using Canonical Correlation Analysis (CCA) to examine associations between groups of variables. Regression analyses, including multiple linear regression, as well as time series-specific methods like cross-correlation functions (CCF) and ARIMA modeling, are also leveraged to model complex interactions and temporal dynamics within the health data.

[0080] Advanced statistical modeling approaches, such as Bayesian Networks and Structural Equation Modeling (SEM), are considered for capturing more sophisticated and potentially causal relationships among variables. These models enable the system to incorporate prior knowledge, handle uncertainty, and infer latent structures or pathways underlying the observed health data. By employing these advanced techniques, the system can generate deeper insights and more robust predictions regarding individual health risks and outcomes. [0081] In certain embodiments, input data, such as the health input data 109 and the general population data 134, can be pre-processed prior to step 230. For example, data may be organized and synchronized by aligning sensor inputs from multiple devices based on their timestamps to ensure that information from different sources is temporally consistent. This process may include adjusting for differences in sampling rates by normalizing the data so that all streams are comparable over time. Temporal interpolation techniques can be applied to fill in missing data points or to synchronize asynchronous signals, enabling accurate analysis and integration of multi-modal health data.

[0082] In another non-limiting example, the input data can be pre-processed by utilizing data cleaning techniques to enhance the quality and reliability of the collected health inputs before further analysis. This step may include detecting and correcting outliers or artifacts that could distort insights, applying denoising algorithms to remove unwanted noise from sensor signals, and normalizing or scaling the data to ensure that values from different sensors are on comparable scales. These procedures help to reduce errors and variability, resulting in cleaner datasets that support more accurate downstream analysis.

[0083] In another non-limiting example, the input data can be pre-processed by using modality specific feature extraction to derive meaningful information tailored to each sensor type. For example, heart rate variability might be extracted from ECG signals, hydration patterns may be detected from camera data, or specific chemical markers from smell sensors. By identifying and isolating relevant features unique to each modality, the system can generate richer and more actionable health and wellness insights for the user.

[0084] The method 200 includes a step 240 of determining when the one or more relationships indicate the individual is experiencing a health event. As described above, the step 240 of determining when the relationships indicate the individual is experiencing a health event may be performed by using a variety of methods, including, by non-limiting example, aggregation of collected data from the one or more input devices 102, data benchmarking of the health input data 109 with the general population data 134 to identify differences, similarities, and deviations from a general population, statistical analysis that analyzes a relationship between the health input data 109 and the general population data 134 (e.g., the use of regression analysis, a calculation of correlation coefficients, or the testing of various hypotheses within the data set), an identification of trends or deviations from a norm for one or more forms of health input data 109, and an interpretation of an individual’s overall health and wellness trends through the extraction of health insights that can allow a healthcare professional to make health recommendations. Furthermore, in certain embodiments, the step 240 of determining when the relationships indicate a health event may be performed by an artificial intelligence component 148.

[0085] When the computing device 140 determines the relationship indicates a person may be experiencing a health event, it may perform a step 250 of generating an alert to be transmitted back to the individual via the network 120. In certain examples, the alert can be transmitted to the mobile device 110 assigned to the individual.

[0086] When the computing device 140 determines the relationship does not indicate an individual is experiencing a health event, it may repeat steps 210, 220, 230, and 240 to reassess any risks for experiencing a health event. In certain embodiments, the reassessment of steps 210, 220, 230, and 240 is performed automatically after a predefined time period since the last assessment. In certain embodiments, the reassessment of steps 210, 220, 230, and 240 can be manually requested by the individual. However, where the relationship does not indicate a person may experience a health event, the relationship may nevertheless provide helpful information for the user regarding beneficial health events.

[0087] FIG. 4 is a schematic diagram of an example system of sensors 106 for collecting at least one health input 109. [0088] The system of sensors 106 can include, by non-limiting example, a temperature sensor 302, a weight sensor 304, a stool analysis sensor 306, a urine analysis sensor 308, and an environment sensor 310, and a biometric sensor 312. Each sensor 106 measures one or more numerical values that can indicate a person may experience a health event if the one or more numerical values are outside of an accepted range of values within the general population data 134. The temperature sensor 302 measures the internal body temperature of the individual. The weight sensor 304 measures a weight of the individual. Furthermore, the stool analysis sensor 306 may analyze characteristics of a person’s stool (feces) after it has been deposited in the toilet. The stool sensor may use a variety of technologies, such as optical sensors (including cameras or spectrometers), chemical sensors, or other analytical devices, to detect and measure parameters like color, consistency, presence of blood, biomarkers, or other chemical components in the stool. The data collected by these smell sensors can be integrated with other health inputs to provide additional context for assessing an individual’s wellness, supporting early detection of potential health issues, and generating more comprehensive and personalized health insights.

[0089] The urine analysis sensor 308 analyzes the contents of urine from the individual, which may include, by non-limiting example, a pH value of the urine, a specific gravity of the urine, and a presence of protein, glucose, ketones, blood, bilirubin, leukocytes, nitrites, or other components that may be found within urine. The environment sensor 310 measures various environmental aspects surrounding the individual, such as an air temperature, an air quality, and other factors. The biometric sensor 312 measures various biometric traits of a person, which may include, by non-limiting example, vital signs such as a heart rate and breathing rate.

[0090] In certain embodiments, the sensors 106 include a photoplethysmography (PPG) sensor 314, which is an optical sensor that measures volumetric changes in blood circulation by emitting light into the skin and detecting the amount of light either transmitted or reflected to a photo detector. In the context of the present system, the PPG sensor 314 may be used to non-invasively collect health input data such as heart rate and blood oxygen saturation from the user. This data can be continuously captured and transmitted to one or more devices to be stored, analyzed, and compared to baseline or population data to generate individual health and wellness insights and evaluate health and wellness trends.

[0091] In some embodiments, the sensors 106 include an electrocardiogram (ECG) sensor 316, which is a device that detects and records the electrical activity generated by the heart as it contracts and relaxes. In the present system, the ECG sensor 316 is used to capture detailed cardiac data such as heart rate, heart rhythm, and the presence of arrhythmias or other cardiac indicators. These measurements can be transmitted to the systems described herein to enable continuous or periodic monitoring of cardiac health, which can be analyzed alongside other health data to provide personalized wellness insights and assess health and wellness trends. [0092] In other embodiments, the sensors 106 include a camera 318 that may be integrated into the system to visually capture images or video of the user or their environment. In this application, the camera can be used for a variety of purposes, such as recording color and quality of patient emissions. The visual data collected can be processed and correlated with other health inputs to enhance the accuracy and personalization of health and wellness insights delivered to the user.

[0093] In certain embodiments, the sensors 106 include a microphone 320 for capturing sound from the user or their surroundings. Within the present system, the microphone 320 may be used to collect audio inputs which may include, by non-limiting example, respiratory sounds, speech patterns, coughs, or other sounds emitted by an individual. Furthermore, the microphone 320 may be used to measure urine flow, which may indicate one or more metrics for predicting prostate health in men and pelvic floor health in women. Audio data collected by the microphone 320 can be analyzed to extract relevant health indicators, which are then integrated with other sensor data to provide a comprehensive assessment of the user’s wellbeing and potential health risks.

[0094] In certain embodiments, the sensors 106 include a radar device 322 that utilizes radio waves to detect motion, measure vital signs, and monitor the presence or activity of the user without direct contact. In the context of the systems described herein, the radar device 322 can be employed to track respiration rate, heart rate, or movement patterns, even through clothing or bedding. The non-invasive data collected by the radar device 322 is transmitted to the system for analysis, contributing additional physiological information that supports the generation of individualized health insights and health and wellness trends.

[0095] In certain embodiments, the sensors 106 may include other sensors 324. The other sensors 324 may include, by non-limiting example, devices such as electronic noses (E- noses), volatile organic compound (VOC) sensors, blood pressure sensors, or other smell sensors designed to detect and analyze chemical signatures in the environment or bodily emissions. In the context of this system, these sensors can be used to monitor air quality, identify the presence of specific gases or odors, or analyze breath and waste for biomarkers associated with certain health conditions. The blood pressure sensor measures a blood pressure of the individual (which is generally measured in two values including a systolic and diastolic blood pressure).

[0096] FIG. 5 is a perspective view of a toilet seat 500 having a plurality of the sensors 106 shown in FIG. 4, according to at least one embodiment of the present disclosure. As illustrated and described above, the toilet seat 500 may include any combination of the one or more sensors 106 such as the temperature sensor 302, the weight sensor 304, the blood pressure sensor 306, the urine analysis sensor 308, the environ ental sensor 310, the biometric sensor 312, the PPG sensor 314, the ECG sensor 316, the camera 318, the microphone 320, the radar device 322, or other sensors 324. Although example placements for the sensors 106 along the toilet seat 500 are shown, these placements are not considered limiting, and the sensors 106 may be positioned at any optimal location to allow for their functionalities to be performed.

[0097] FIG. 6 is a method flow diagram 600 of a method for training a machine learning model, according to at least one embodiment of the present disclosure.

[0098] The method 600 includes a step 610 of collecting health input data 109 from a variety of sensors 106 and devices, including wearable technology, mobile applications, and specialized apparatus such as sensor-equipped toilet seats. These input devices gather a wide range of health metrics as illustrated and described above with reference to FIGS. 1-5, including, by non-limiting example, heart rate, blood pressure, oxygen saturation, hydration status, and stool characteristics, as well as self-reported data including exercise logs, mood, sleep, and nutrition. The collected data is transmitted over a network and stored in a memory structure, which may also incorporate general population data and an individual’s historic health records. In certain embodiments, this comprehensive and continuous data acquisition is used in subsequent analysis and model development.

[0099] The method 600 includes a step 620 of pre-processing data to ensure quality, consistency, and usability for model training. Preprocessing steps are illustrated and described above with reference to FIGS. 1-3 and may include, by non-limiting example, synchronizing data streams from multiple sensors 106 based on timestamps, cleaning the data to remove noise and outliers, and handling missing values through interpolation or imputation. Additionally, modality-specific feature extraction is performed to derive meaningful variables from raw sensor signals, such as extracting heart rate variability from ECG data or hydration status from urine analysis sensors. This curated dataset is then normalized and scaled, resulting in a structured and reliable input for machine learning algorithms. [0100] The method 600 includes a step 630 of training a machine learning model. Health input data 109 (such as the health input data that is curated using pre-processing step 620) is used to train a machine learning model designed to identify relationships between individual health metrics and general population or individual historical and real-time health data. Supervised learning methods may be employed, leveraging labeled datasets to teach the model how to classify health states or predict when a person may experience a health event. Alternatively, unsupervised techniques such as clustering and dimensionality reduction may be used to uncover hidden patterns within the data. The training phase involves iterative optimization of model parameters to maximize predictive accuracy and robustness in detecting health risks.

[0101] The method 600 includes a step 640 of evaluating the model to assess its performance in predicting health and wellness trends and generating wellness insights. In certain embodiments, evaluation metrics such as accuracy, sensitivity, specificity, and area under the ROC curve are calculated using a separate validation dataset. The model's ability to generalize to new, unseen data is tested, and its predictions are compared against known outcomes or expert assessments. In certain embodiments, evaluation ensures that the model meets predefined standards for reliability, safety, and healthcare relevance before deployment. [0102] The method 600 includes a step 650 of determining whether performance criteria was met. This step involves determining whether the trained model satisfies the established performance criteria. In certain embodiments, these criteria may include thresholds for predictive accuracy, false positive and negative rates, and other domain-specific requirements. If the model meets or exceeds these benchmarks, it is deemed suitable for deployment. If not, the process may transition to a feedback and refinement phase, where shortcomings are analyzed and addressed before retraining.

[0103] The method 600 includes a step 660 of deploying the model if the evaluation of the trained model demonstrates that the predefined performance criteria have been met. In certain embodiments, deployment involves integrating the model into the health and wellness systems described herein to begin processing real -world health input data 109, generating individual risk assessments, and providing actionable insights or alerts to end-users or healthcare providers.

[0104] The method 600 includes a step 670 of providing feedback for refinement if performance criteria is not met. This feedback may include, by non-limiting example, feedback from a user as illustrated and described with reference to FIGS.8- 10 below. Furthermore, in certain embodiments, this feedback may include information on which specific metrics or aspects of the model's predictions were inadequate, such as low accuracy, poor sensitivity to certain health events, or high rates of false positives or negatives. The feedback may be used to identify weaknesses or limitations in the current iteration of the model, informing the next steps in the improvement process.

[0105] The method 600 includes a step 680 of refining the machine learning model to improve its performance. This refinement may involve adjusting model parameters, incorporating additional features or data sources, selecting different machine learning techniques, or addressing issues such as data imbalance or overfitting. The goal of this step is to enhance the model’s ability to accurately identify relationships and predict health and wellness trends, thereby increasing its utility and reliability in real-world applications.

[0106] The method 600 includes a step 690 of retraining the machine learning model and reverting back to step 630 to learn from the improvements and corrections made during the refinement process. Once retraining is complete, the process loops back to the model evaluation step (step 630), where the model’s performance is reassessed. This iterative cycle continues until the model consistently meets the established performance criteria and is ready for deployment.

[0107] FIG. 7 is a method flow diagram of a method 700 for generating individual health and wellness insights, according to at least one embodiment of the present disclosure.

[0108] The method 700 includes a step 710 of collecting health input data 109 from a variety of sources as illustrated and described above with reference to FIGS. 1-6. The data encompasses a wide range of health metrics. Furthermore the data encompasses baseline individual historical and real-time health data and benchmark data that can be used to compare individual health metrics to in order to identify any relationships in user health data that warrant further action.

[0109] The method 700 includes a step 720 of storing the health input data 109. Once collected, the health input data 109 is securely stored in a memory structure, which may also include general population data and the individual’s historical health records. This centralized storage ensures that the data is readily accessible for subsequent processing and analysis, and supports the integration of multiple data types from different sources. This step is illustrated and described above in further detail with respect to FIGS. 1-6.

[0110] The method 700 includes a step 730 of processing the health input data 109 to ensure quality and consistency. This includes synchronizing data streams based on timestamps, cleaning the data to remove noise and outliers, handling missing values, and performing modality-specific feature extraction. These preprocessing steps are critical for generating a reliable and structured dataset that supports accurate analysis. This step is illustrated and described above in further detail with respect to FIGS. 1-6.

[OHl] The method 700 includes a step 740 of analyzing relationships between the health input data 109 and the individual’s historic health data, benchmark health data, or both. This analysis may involve statistical techniques such as regression analysis, correlation coefficients, and machine learning methods, as previously described above with reference to FIGS. 1-6.

[0112] The method 700 includes a step 750 of determining whether the identified relationships are statistically significant. In certain embodiments, this is accomplished using one or more statistical tests and thresholds, such as calculating p-values, confidence intervals, or comparing correlation coefficients against predefined cutoffs. Systems and methods for identifying whether the relationships are significant using statistical tests and thresholds are illustrated and described above in further detail with reference to FIGS. 1-6. Significance is further assessed by evaluating whether deviations from baseline or population norms are large enough to indicate a potential health risk.

[0113] The method 700 includes a step 760 of generating an alert if significant relationships are detected. In certain embodiments, the alert may include specific information about the health condition or risk identified, as well as relevant background such as the user’s medical history or recommended next steps. The generation of the alert is illustrated and described above in further detail with respect to FIGS. 1-6.

[0114] The method 700 includes a step 770 of providing a personalized health or wellness insight based on the analysis. This insight may include recommendations, trend summaries, or feedback on positive health behaviors, helping users to better understand their health status and make informed decisions. The generation of a personalized health and wellness insight is illustrated and described above with reference to FIGS. 1-6.

[0115] The method 700 includes a step 780 of ending the process if no significant relationship is detected or after a health and wellness insight is provided.

[0116] FIG. 8 is a method flow diagram of a method 800 for a user to use a health and wellness system, according to at least one embodiment of the present disclosure.

[0117] The method 800 includes a step 810 of providing sample data to the system. This typically occurs when the user sits on the sensor-equipped toilet seat, (such as the system illustrated and described above with respect to FIGS. 1-7) which collects a range of health input data 109 as illustrated and described above with reference to FIGS. 1-7. [0118] The method 800 includes a step 820 of receiving an output from the system. The output consists of health and wellness insights generated by the system, which may include alerts about potential health and wellness trends, personalized recommendations, or summaries of the user’s health status. The output is typically delivered through a user interface, such as a mobile application or display device. Examples of user interfaces 900, 1000 are illustrated and described in further detail below with reference to FIGS. 9-10.

[0119] The method 800 includes a step 830 of a user reviewing the output from the system. This review allows the user to understand the insights, alerts, or recommendations generated based on their sample data. The output may include explanations of detected trends, identified risks, or suggestions for improving health and wellness, empowering the user to make informed decisions about their health.

[0120] The method 800 includes a step 840 of a user providing feedback to the system. This feedback may address the accuracy or relevance of the insights, report any discrepancies, or supply additional contextual information (such as recent dietary or lifestyle changes) that could further refine future outputs. In certain embodiments, user feedback is desirable for continuously improving the system’s predictive accuracy and personalization.

[0121] The method 900 includes a step 850 of a user adjusting their behavior based on the output generated by the system. In certain embodiments, adjustments to a user’s behavior may include making changes to diet, exercise, hydration, medication adherence, or other lifestyle factors as suggested by the system’s analysis. By acting on these personalized insights, users can proactively manage their health and analyze health and wellness trends. [0122] FIG. 9 is a schematic view of a user interface 900, according to at least one embodiment of the present disclosure. The user interface includes a menu icon 910, a selfreport session icon 920, a date 930 where one or more health and wellness insights were generated, one or more categories for health and wellness 940, 950, 960, a manual start icon 970 for manually starting a recording session to record sample, and one or more other icons for accessing additional information, including a home page icon 980, one or more icons corresponding to additional information for the health and wellness categories (shown as H&W 2, H&W 2, and H&W 3) 984, 986, 988, and a timeline icon 990 for accessing individual historic health data recorded over time.

[0123] The menu icon 910 serves as a primary navigation tool within the user interface. When selected, it typically opens a sidebar or dropdown menu, providing the user with access to various system settings, account management options, and additional application features. This centralized access point enables users to efficiently navigate between different sections of the health and wellness platform, customize their experience, and manage notification preferences or device connections.

[0124] The self-report session icon 920 allows users to initiate a self-reporting session, where they can manually input information that may not be automatically captured by the system’s sensors. This can include subjective data such as mood, symptoms, dietary intake, or physical activity. By enabling self-reporting, the system can integrate both objective sensor data and valuable user-provided context, resulting in more comprehensive and personalized health and wellness insights.

[0125] The one or more categories for health and wellness 940, 950, 960 may relate to any metric for health and wellness measured by the system. For each of the one or more categories for health and wellness 940, 950, 960, a user may provide user feedback 942, 952, 962 to the system that may not be immediately detectable by the one or more sensors 106. Furthermore, each of the one or more categories for health and wellness 940, 950, 960 includes a timestamp 944, 954, 964 indicating when the measurement was taken.

[0126] The manual start icon 970 enables users to manually initiate a new recording session for sample collection. This feature is particularly useful if the user wishes to capture data outside of the system’s automatic collection schedule, such as after experiencing specific symptoms or engaging in unusual activities. By providing control over sample timing, the system supports more flexible and user-driven monitoring.

[0127] The home page icon 980 provides a quick way for users to return to the main dashboard or landing screen of the application. From the home page, users can access a summary of their overall health status, recent alerts, and key insights. This central hub streamlines navigation and ensures users can easily orient themselves within the application. [0128] In certain embodiments, the one or more icons correspond to additional information for the health and wellness categories 984, 986, 988.

[0129] The timeline icon 990 allows users to access a chronological view of their individual historic health data. Through this feature, users can review trends, patterns, and changes in their health metrics over extended periods. The timeline visualization facilitates longitudinal analysis, helping users and healthcare providers to monitor progress, identify recurring issues, and evaluate the effectiveness of behavior changes or interventions.

[0130] FIG. 10 is another embodiment of a schematic view of a user interface 1000, according to at least one embodiment of the present disclosure. The user interface includes a user feedback menu 1010 for providing user feedback 1030. The user feedback 1030 may include providing one or more tags 1030, including the tags shown or a custom tag, may be added to the system using a create custom tag icon 1020. The tags 1030 may be saved to the system using a save tags icon 1040. In certain embodiments, the tags relate to one or more symptoms an individual is experiencing.

[0131] It should be understood that the present disclosure is not limited to the details or methodology set forth in the description or illustrated in the figures. It should also be understood that the terminology used herein is for the purpose of description only and should not be regarded as limiting.

[0132] It should be understood that various aspects disclosed herein may be combined in different combinations than the combinations specifically presented in the description and accompanying drawings. It should also be understood that, depending on the example, certain acts or events of any of the processes or methods described herein may be performed in a different sequence, may be added, merged, or left out altogether (e.g., all described acts or events may not be necessary to carry out the techniques). In addition, while certain aspects of this disclosure are described as being performed by a single module or unit for purposes of clarity, it should be understood that the techniques of this disclosure may be performed by a combination of units or modules.

[0133] In one or more examples, the described techniques may be implemented in hardware, software, firmware, or any combination thereof.

Claims

CLAIMS What is claimed is:

1. A system comprising: one or more input devices for collecting at least one health input from the individual and transmit the at least one health input to a computing device via a network; the computing device for receiving the at least one health input from the individual, the computing device comprising: a processor; and a device comprising a non-transitory storage medium encoded with instructions executable by the processor which, when executed by the processor, cause the processor to: store the at least one health input in a device memory, wherein the device memory comprises general population data from a population of individuals and historical health data for an individual; identify one or more relationships between the at least one health input and the general population data or the historical health data; determine when the one or more relationships indicate the individual may experience a health event; and generate an alert when the one or more relationships indicate the individual may experience a health event.

2. The system of claim 1, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: transmit the alert to a mobile device assigned to the individual.

3. The system of claim 1, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: request a manual input from the individual relating to one or more factors that affect the at least one health input of the individual.

4. The system of claim 3, wherein the one or more factors comprise at least one of a weather report, an exercise log, a mood log, a sleep log, and a nutrition log.

5. The system of claim 1, wherein the one or more input devices comprise at least one sensor.

6. The system of claim 5, wherein the at least one sensor comprises at least one of a temperature sensor, a weight sensor, a pressure sensor, a urine analysis sensor, and an environment sensor, and a biometric sensor a PPG sensor, an ECG sensor, a camera, a microphone, a radar device, a smell sensor, or a stool sensor.

7. The system of claim 1, wherein the computing device receives at least one manual input from a user.

8. The system of claim 7, wherein the manual input includes information captured by the mobile device assigned to the user.

9. The system of claim 1, wherein the device comprises the non-transitory storage medium encoded with the instructions executable by the processor which, when executed by the processor, further cause the processor to: perform one or more pre-processing operations to curate the at least one health input, wherein the one or more pre-processing operations include at least one of synchronizing the at least one health input based on timestamp alignment, performing one or more data cleaning operations, or performing one or more modality specific feature extractions.

10. The system of claim 1, wherein: the one or more relationships between the at least one health input and the general population data or the historical health data are generated using one or more statistical analysis operations; and the one or more statistical analysis operations include at least one of a univariate statistical measurement, a bivariate correlation technique, a multivariate statistical technique, or an advanced statistical model.

11. The system of claim 10, wherein the bivariate correlation technique includes at least one of a Pearson correlation coefficient, a Spearman's rank correlation, a Kendall's Tau, or scatter plots.

12. The system of claim 1, wherein the multivariate statistical technique includes at least one of: a correlation matrix or heatmap, a Principal Component Analysis, a Factor Analysis, a Canonical Correlation Analysis, a regression analysis including multiple linear regression, or a time series specific statistical methods including a cross-correlation function and an autoregressive integrated moving average.

13. A method comprising: receiving at least one health input from one or more input devices via a network; storing the at least one health input in a device memory, wherein the device memory comprises a general population data from a population of individuals and historical health data from a user; identifying one or more relationships between the at least one health input and the general population data or the historical health data; determining when the one or more relationships indicate the individual may experience a health event; and generating an alert when the one or more relationships indicate the individual may experience a health event.

14. The method of claim 13, further comprising transmitting the alert to a mobile device assigned to the individual.

15. The method of claim 13, further comprising receiving at least one manual input from a user.

16. The method of claim 13, further comprising generating the one or more relationships between the at least one health input and the general population data or the historical health data using one or more statistical analysis operations, wherein the one or more statistical analysis operations include at least one of a univariate statistical measurement, a bivariate correlation technique, a multivariate statistical technique, or an advanced statistical model.

17. The method of claim 16, wherein the bivariate correlation technique includes at least one of a Pearson correlation coefficient, a Spearman's rank correlation, a Kendall's Tau, or scatter plots.

18. The method of claim 16, wherein the multivariate statistical technique includes at least one of: a correlation matrix or heatmap, a Principal Component Analysis, a Factor Analysis, a Canonical Correlation Analysis, a regression analysis including multiple linear regression, or a time series specific statistical methods including a cross-correlation function and an autoregressive integrated moving average.

19. A method for training a machine learning model to generate one or more health and wellness insights, comprising: providing health input data comprising general population data from a population of individuals and historical health data from a user; performing one or more pre-processing steps to curate the health input data; identifying one or more relationships between the at least one health input and the general population data or the historical health data; determining when the one or more relationships indicate the individual may experience a health event; determining whether the one or more relationships exceed a predetermined threshold value; and locking the machine learning model.

20. The method of claim 19, further comprising providing one or more labels to train the machine learning model using supervised learning.