Analysis device, analysis system, analysis method, and program

Abstract

Provided is an analysis device and the like that achieves improved prediction accuracy even when training data exhibiting large variability among data points are used.　The analysis device comprises: a first acquisition unit that acquires first data relating to characteristics of samples; a second acquisition unit that acquires second data relating to characteristics of the samples; a learning unit that generates a trained model on the basis of sample data groups including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for the known samples; and an analysis unit that inputs the first data acquired by the first acquisition unit for an unknown sample to the trained model generated by the learning unit and outputs the result of analysis on the unknown sample. The learning unit tests for significant differences between the sample data groups by a statistical hypothesis test method, selects combinations of the subject data groups determined to have a significant difference, and generates training data.

Description

Analysis Device, Analysis System, Analysis Method, and Program

[0001] The present disclosure relates to an analysis device, an analysis system, an analysis method, and a program.

[0002] It is known to estimate sensory evaluations such as smell, taste, impression, and inference for a subject from the characteristics of the subject. For example, in the technique described in Patent Document 1, taste evaluation for food and drink is estimated using a taste estimation model created by machine learning.

[0003] Japanese Patent Application Laid-Open No. 2022-047369

[0004] Generally, if there is a large variation between the learning data used in machine learning, the accuracy of the created machine learning model will be low. In particular, when the results of sensory evaluation are used as learning data, the results are likely to vary greatly depending on the panelists who evaluate. Therefore, it is difficult to perform analysis of sensory evaluation by machine learning with high prediction accuracy, and further improvement is required.

[0005] The problem to be solved by the present disclosure is to provide an analysis device or the like that improves the prediction accuracy when using learning data with a large variation between data.

[0006] To solve the above problems, the analysis device of the present disclosure includes: a first acquisition unit that acquires first data about the characteristics of a subject; a second acquisition unit that acquires second data about the characteristics of the subject; a learning unit that generates a learned model based on a subject data group including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for the known subject; and an analysis unit that inputs the first data acquired by the first acquisition unit for the unknown subject into the learned model generated by the learning unit and outputs an analysis result for the unknown subject. The learning unit tests the significant difference between the subject data groups by a statistical hypothesis testing method, and selects a combination of the subject data groups for which a significant difference is determined to generate learning data.

[0007] The analysis system of this disclosure comprises the analysis device and the measuring device for measuring the first data, wherein the first acquisition unit acquires the first data measured by the measuring device.

[0008] The analysis method of this disclosure is an analysis method performed by an analysis device, comprising: a first acquisition step of acquiring first data about the characteristics of a subject; a second acquisition step of acquiring second data about the characteristics of the subject; a learning step of generating a trained model based on a group of subject data including the first data acquired in the first acquisition step and the second data acquired in the second acquisition step for a known subject; and an analysis step of inputting the first data acquired in the first acquisition step for an unknown subject into the trained model generated by the learning step and outputting an analysis result for the unknown subject, wherein the learning step tests for significant differences between the groups of subject data using a statistical hypothesis testing method, selects combinations of the groups of subject data that are determined to have significant differences and generates training data.

[0009] The program of this disclosure is a program that causes the computer of an analysis device to function as a first acquisition unit that acquires first data about the characteristics of a subject; a second acquisition unit that acquires second data about the characteristics of the subject; a learning unit that generates a trained model based on a group of subject data including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for a known subject; and an analysis unit that inputs the first data acquired by the first acquisition unit for an unknown subject into the trained model generated by the learning unit and outputs analysis results for the unknown subject, wherein the learning unit tests for significant differences between the groups of subject data using statistical hypothesis testing methods, selects combinations of the groups of subject data that are determined to have significant differences and generates training data.

[0010] According to this disclosure, it is possible to improve the predictive accuracy of analysis when using training data that has large variability among the data.

[0011] This is a block diagram illustrating the schematic configuration of the analysis system. This is a schematic cross-sectional diagram of a laminate. This is a schematic cross-sectional diagram of a laminate coated with a sample. This is a schematic cross-sectional diagram of a laminate coated with a light-emitting probe. This is a schematic cross-sectional diagram of an organic EL element. This is a diagram illustrating the mechanism by which the target substance and light-emitting dye molecules interact to produce light. This is a diagram illustrating the synthesis method of the light-emitting dye molecules. This shows a flowchart of the measurement method. This is a flowchart illustrating an example of the operation of the analysis system. This is a flowchart illustrating an example of the operation of the learning process in the analysis system.

[0012] Hereinafter, one or more embodiments of this disclosure will be described with reference to the drawings. However, the scope of this disclosure is not limited to the disclosed embodiments.

[0013] [Configuration of the Analysis System] Figure 1 is a block diagram showing the schematic configuration of the analysis system 100 of this embodiment. The analysis system 100 comprises a measuring device 10 and an analysis device 20. The measuring device 10 and the analysis device 20 may communicate with each other. Alternatively, the measuring device 10 and the analysis device 20 may be integrated. The measuring device 10 comprises a measuring unit 11, a communication unit 12, and a control unit 13. The analysis device 20 comprises a display unit 21, an operation unit 22, a communication unit 23, and a control unit 24.

[0014] In this embodiment, the first data and the second data are not particularly limited as long as they are data relating to the characteristics of the subject. However, from the viewpoint of obtaining the effects of this disclosure significantly, it is preferable that at least the first data is a measured value obtained by measuring a physical quantity relating to the characteristics of the subject. The first data can be measured by the measuring device 10.

[0015] The second data may be measured values ​​obtained by measuring physical quantities related to the characteristics of the subject, or it may be values ​​input by the user of the analysis system 100, such as evaluation values ​​obtained through sensory evaluation. In this embodiment, "sensory evaluation" refers to the evaluation of the characteristics, functions, sensations, etc. of the subject using a person's five senses, and the results of that evaluation. Examples of evaluation items for sensory evaluation include flavor, texture, smell, and feel.

[0016] As mentioned earlier, machine learning analysis based on datasets with significant variability tends to result in lower prediction accuracy. For example, in sensory evaluation, evaluations by panelists tend to vary greatly. Furthermore, panelists do not always use consistent criteria for evaluation. Therefore, analyzing sensory evaluation using machine learning is extremely difficult.

[0017] In this embodiment, in a dataset with significant variability, the presence or absence of statistically significant differences between datasets is determined by a test. Combinations of datasets determined to have a statistically significant difference are selected to generate training data. This allows for analysis with high predictive accuracy for items that generally tend to have large variability between datasets.

[0018] The details of each device are described below.

[0019] [Measurement device] (Measurement unit) The number of data points of the first data measured by the measurement unit 11 and acquired by the control unit 24 of the analysis device 20 is not particularly limited, but the more data points there are, the better the prediction accuracy of the analysis. The number of data points is appropriately selected depending on the measurement method, etc. It is preferable that the number of data points be equal to or greater than the number of explanatory variables. If a large number of data points can be acquired, it is preferable that the number of data points be 10 times or more the number of explanatory variables, and more preferably 100 times or more.

[0020] The physical quantities to be measured are not particularly limited, but from the viewpoint of obtaining a large amount of complex data, it is preferable to measure optical properties that change behavior in interaction with the subject. Other physical quantities include viscosity, density, pH value, etc. The device used to measure the physical quantities is not particularly limited, and any measuring device can be used.

[0021] The following describes the measurement unit 11, which measures optical properties whose behavior changes in response to interaction with the subject.

[0022] One method for measuring such optical properties involves irradiating a sample with light from a light source and obtaining spectral data of the light obtained from the sample. Examples of light emitted from a light source include X-rays, ultraviolet rays, visible light, and infrared rays.

[0023] In this embodiment, the measurement unit 11 preferably uses a light-emitting probe to acquire spectral data of light obtained from the sample. In this embodiment, a "light-emitting probe" refers to a substance that has the property of binding to or associating with the target substance and emits light when returning from an excited state to a ground state. By using a light-emitting probe, accurate data about the target substance can be obtained, and the prediction accuracy of the analysis can be improved.

[0024] If the sample contains multiple substances, the luminescent probe may bind to or associate only with specific target substances within the sample. The luminescent probe may be a molecule, a molecular aggregate, or a particle containing molecules.

[0025] The method for acquiring spectral data of light obtained from a sample using a light-emitting probe is not particularly limited. For example, one method is to use an organic electroluminescent (EL) element. Another method is to use a light-emitting dye molecule having a main chain of a specific structure as the light-emitting probe.

[0026] (Organic EL Elements) The following configurations are possible for organic electroluminescent (EL) elements: (i) Transparent substrate / anode / light-emitting layer (detection region) / cathode (ii) Transparent substrate / anode / light-emitting layer (detection region / electron transport layer) / cathode (iii) Transparent substrate / anode / light-emitting layer (hole transport layer / detection region) / cathode (iv) Transparent substrate / anode / light-emitting layer (hole transport layer / detection region / electron transport layer) / cathode (v) Transparent substrate / anode / light-emitting layer (hole transport layer / detection region / electron transport layer / electron injection layer) / cathode (vi) Transparent substrate / anode / light-emitting layer (hole injection layer / hole transport layer / electron blocking layer / detection region / hole blocking layer / electron transport layer) / cathode

[0027] The above-described organic EL element has a light-emitting layer between its electrodes, and the light-emitting layer is the detection region. The organic EL element can emit light when the test subject and the light-emitting probe interact within the light-emitting layer. Various data can be obtained by acquiring the light emission information and current density of the light-emitting probe. A method for acquiring data using the organic EL element will be explained with reference to Figures 2 to 5.

[0028] In this method, first, a laminate comprising a transparent substrate 111, a transparent electrode 112 (e.g., an ITO film), a hole transport layer 113 (e.g., a polyaromatic diamine), and a receiving layer 114 (e.g., polystyrene) is prepared (Figure 2). The transparent substrate 111, transparent electrode 112, hole transport layer 113, etc., can be the same as those used in known organic EL devices. The receiving layer 114 only needs to be able to receive the subject and the light-emitting probe. The receiving layer 114 may also serve as the host layer.

[0029] Next, the sample layer 121 is applied to a desired area of ​​the receiving layer 114 of the laminate using any method (Figure 3).

[0030] Next, the luminescent probe is applied in a pattern to desired locations on the receiving layer 114 coated with the subject layer 121 using an inkjet method or the like (Figure 4). As a result, the areas coated with the luminescent probe become the luminescent layer 122. For example, by applying a component that dissolves the receiving layer 114 together with the luminescent probe, the applied luminescent probe, a part of the subject layer 121, and a part of the receiving layer 114 are mixed, and an integrated luminescent layer 122 is formed. In the luminescent layer 122, the subject and the luminescent probe interact. Although not shown in the figure, the luminescent probe may also be applied to areas where the subject layer 121 is not coated to obtain luminescence information from the luminescent probe alone.

[0031] Finally, a counter electrode layer 115, which is paired with the transparent electrode 112, is placed on the light-emitting layer 122 to fabricate an organic EL element (Figure 5).

[0032] In the fabricated organic EL element, the test subject and the light-emitting probe in the light-emitting layer 122 are excited by a conventional method, and emission spectral data is acquired.

[0033] When acquiring a light spectrum using an organic EL element, i.e., when using an organic EL element as a sensing device, fluorescent compounds, delayed-fluorescence compounds, and phosphorescent compounds can be used as light-emitting probes. Furthermore, different phosphorescent compounds may be used in combination as light-emitting probes, or phosphorescent compounds and fluorescent compounds may be used in combination. This allows for the acquisition of any desired emission color. Additionally, multiple light-emitting compounds with different emission colors may be combined to produce white light.

[0034] In this specification, "fluorescent compound" refers to a compound that emits fluorescence other than delayed fluorescence. "Fluorescence" refers to the light emitted when returning from a singlet excited state to the ground state. "Fluorescence other than delayed fluorescence" refers to fluorescence excluding the "delayed fluorescence" exemplified below. Examples of "delayed fluorescence" include "thermally activated delayed fluorescence (TADF)". Examples of "delayed fluorescence" include "triplet-triplet annihilation (TTA) delayed fluorescence".

[0035] In other words, in this specification, "fluorescent compound" does not include "delayed fluorescent compounds" such as "thermally activated delayed fluorescent compounds" and "triplet-triplet annihilation delayed fluorescent compounds." "Fluorescent compound" refers to a compound that does not undergo upconversion by reverse intersystem crossing from the lowest excited triplet energy level to the lowest excited singlet energy level.

[0036] Fluorescent compounds do not necessarily need to be heavy metal complexes like phosphorescent compounds. So-called organic compounds, composed of common combinations of elements such as carbon, oxygen, nitrogen, and hydrogen, can be used as fluorescent compounds. Other nonmetallic elements such as phosphorus, sulfur, and silicon may also be used as fluorescent compounds. Complexes of typical metals such as aluminum and zinc may also be used as fluorescent compounds. Known fluorescent compounds used in the light-emitting layers of organic EL devices may also be used as fluorescent compounds.

[0037] In this specification, "phosphorescent compound" refers to a compound that emits phosphorescence. Specifically, a "phosphorescent compound" refers to a compound that emits phosphorescence at room temperature (25°C) and has a phosphorescence quantum yield of 0.01 or higher at 25°C. A phosphorescence quantum yield of 0.1 or higher is preferred.

[0038] "Phosphorescence" refers to the light emitted when returning from a triplet excited state to the ground state. Known phosphorescent compounds used in the light-emitting layers of organic EL devices can be used as the phosphorescent compound.

[0039] In this specification, "delayed fluorescence compound" refers to a compound that emits delayed fluorescence. "Delayed fluorescence" refers to the light emitted when a singlet excited state returns to the ground state as a result of upconversion by reverse intersystem crossing from the lowest excited triplet energy level to the lowest excited singlet energy level. Known delayed fluorescence compounds used in the light-emitting layer of organic EL elements can be used as the delayed fluorescence compound.

[0040] When using an organic EL element as a sensing device, multiple data points may be taken for each of the different emission colors (blue, green, and red) at intervals of a few nanometers. Alternatively, the current density may be measured at 1V intervals. This allows for the acquisition of a large number of data points.

[0041] (Luminescent dye molecules) The method using luminescent dye molecules described below uses microwells that can form multiple reaction fields on a plane. Different types of luminescent dye molecules are prepared in advance in each well, and the sampled subject is injected into each well to obtain emission spectral data. This allows the state of the subject to be converted into data from multiple emission spectral data.

[0042] This method uses a plate in which multiple wells are arranged regularly. In a plate with such wells, the wells (reaction fields) are physically separated by partitions. Therefore, the test substance and luminescent dye molecules are less likely to mix in adjacent reaction fields, making accurate analysis easier.

[0043] The luminescent dye molecule exhibits two or more types of luminescence selected from fluorescence, phosphorescence, excimer luminescence, exciplex luminescence, thermally activated delayed fluorescence, excited state intramolecular proton luminescence, triplet-triplet annihilation delayed fluorescence, twisted intramolecular charge transfer luminescence, and aggregation-induced organic luminescence with respect to a single excitation light.

[0044] The luminescent dye molecule can be used for analyzing the structure, state, etc. of a specific target substance. Specifically, when the luminescent dye molecule and the target substance interact with each other, the structure, electronic state, etc. in the chromophore or lumophore in the luminescent dye molecule change, and a complex luminescence behavior different from that of the single luminescent dye molecule can be obtained.

[0045] For example, as shown in FIG. 6 with respect to a single excitation light, a luminescent dye molecule that exhibits three different types of luminescence, namely fluorescence, phosphorescence, and excimer luminescence, is allowed to interact with a target substance. Due to the interaction between the target substance and the luminescent dye molecule, the processes of generating fluorescence, phosphorescence, and excimer luminescence change respectively, and the wavelength or lifetime of each light changes. As a result, depending on the structure, state, etc. of the target substance, a large number of complex data combined with these lights can be obtained, and from this data, the structure, state, etc. of the target substance can be grasped very in detail.

[0046] Hereinafter, the specific structure of the luminescent dye molecule according to the present embodiment will be described.

[0047] The main chain of the luminescent dye molecule may have one or more structural units including a sugar structure derived from pentose or hexose and a phosphate ester bond bonded to the sugar structure. The main chain may contain only one of the above structural units or may contain a plurality of them. That is, it may be a structure having one sugar structure and one phosphate ester bond bonded to the sugar structure, or may be a structure containing the sugar structure and the phosphate ester bond alternately.

[0048] Both ends of the main chain of the luminescent dye molecule are sugar structures, and the number of sugar structures is one more than the number of phosphate ester bonds. When the main chain contains a plurality of structural units, the plurality of structural units may be the same as each other or may be different.

[0049] The number of structural units contained in the main chain of the luminescent dye molecule is appropriately selected according to the type of target substance, and is preferably in the range of 2 to 6. If the amount of structural units is large, the luminescent dye molecule is more likely to act specifically on the target substance. In this embodiment, it is preferable to obtain a large amount of data by having the luminescent dye molecule interact with various positions on the target substance. For this reason, it is preferable that the luminescent dye molecule and the target substance have a moderate degree of specificity, not excessively so, and the number of structural units is preferably 6 or less.

[0050] The main chain of the luminescent dye molecule may include structures other than the above-mentioned structural units, to the extent that it does not impair the purpose and effect of this embodiment. The structures at both ends of the main chain are not particularly limited and may include, for example, hydroxyl groups, alkoxy groups, etc.

[0051] Examples of pentoses include ribose, deoxyribose, and xylose. Examples of hexoses include allose, glucose, and mannose. Among these, it is particularly preferable that the sugar structure is derived from ribose or deoxyribose. This allows the main chain of the luminescent pigment molecule to have a structure similar to that of DNA or RNA, and the luminescent pigment molecule to readily interact with DNA or RNA.

[0052] When the above structural unit includes a structure derived from ribose or deoxyribose, the phosphate ester bond is preferably bonded to the carbon at position 3 and position 5 of ribose or deoxyribose. The chromophore or luminescent phore is preferably bonded to the carbon at position 1 of ribose or deoxyribose. In other words, the luminescent dye molecule according to this embodiment preferably includes a structure represented by the following general formula (1a) or (1b).

[0053]

[0054] In the general formulas (1a) and (1b), Y represents a chromophore or luminescent phose as described below.

[0055] The chromophore or luminescent phore of the luminescent dye molecule may have a structure that emits a predetermined type of light independently in response to a single excitation light, or emits a predetermined light through the action of multiple chromophores or luminescent phores. Preferably, the chromophore or luminescent phore is bonded to the sugar structure of the main chain such that the sugar structure is in its β-form.

[0056] In this embodiment, "chromophore" refers to a structure that absorbs light with a wavelength of 300 nm or more, and "luminescent phore" refers to a structure that absorbs light with a wavelength of 300 nm or more and emits light.

[0057] The number of chromophores or luminescent groups in a luminescent dye molecule may be as few as one, provided that the luminescent dye molecule is capable of exhibiting multiple types of luminescence. However, from the viewpoint of making it easier for the luminescent dye molecule to exhibit multiple types of luminescence, the number of chromophores or luminescent groups is preferably two or more, and more preferably in the range of three to six. When a luminescent dye molecule has multiple chromophores or luminescent groups, there may be only one type, or there may be two or more types.

[0058] In a luminescent dye molecule, it is preferable that one chromophore or luminescent phore is bound to one sugar structure in the main chain. Therefore, if a luminescent dye molecule has two or more chromophores or luminescent phore, it is preferable that there are also two or more sugar structures in the main chain. In other words, it is preferable that the number of chromophores or luminescent phore in a luminescent dye molecule is equal to or less than the number of sugar structures in the main chain.

[0059] If the number of chromophores or luminescent groups in a luminescent dye molecule is less than the number of sugar structures in the main chain, some sugar structures will not have chromophores or luminescent groups attached. Sugar structures that do not have chromophores or luminescent groups attached do not need to have other atomic groups attached, but they may have native bases attached to them to the extent that it does not impair the purpose and effect of this embodiment. In this specification, "native bases" refers to adenine, guanine, cytosine, thymine, and uracil.

[0060] The total number of native bases bound to the sugar structure is preferably 50% or less, and more preferably 25% or less, of the total number of sugar structures in the main chain. Having 50% or less native bases suppresses the association of luminescent dye molecules, making the interaction between the test subject and the luminescent dye molecules more dominant. Furthermore, it is preferable that the native and non-native bases are bound in such a way that the sugar structure is in the β-form.

[0061] Examples of fluorescent chromophores or luminescent phoses include structures derived from fluorescein, rhodamine, boron dipyromethene, etc. Examples of phosphorescent chromophores or luminescent phoses include structures derived from iridium complexes, platinum complexes, etc. Examples of excimer-luminescent chromophores or luminescent phoses include structures derived from pyrene, anthracene, perylene, etc. Examples of exciplex-luminescent chromophores or luminescent phoses include structures derived from pyrene-dimethylaniline, etc.

[0062] Examples of chromophores or luminescent molecules that emit thermally activated delayed fluorescence include structures derived from 4CzIPN, DABNA, etc. Examples of chromophores or luminescent molecules that emit excited-state intramolecular proton emission include structures derived from hydroxyphenylbenzoxazole, etc. Examples of chromophores or luminescent molecules that emit triplet-triplet annihilation delayed fluorescence include structures derived from 9,10-diphenylanthracene, rubrene, etc. Examples of chromophores or luminescent molecules that emit twisted intramolecular charge transfer emission include structures derived from diaminoanthracene, diaminonaphthalene, etc. Examples of chromophores or luminescent molecules that emit aggregated organic emission include structures derived from tetraphenylethene, hexaphenylsilole, etc.

[0063] In particular, the chromophore or luminescent phore preferably contains at least one structure selected from a structure that emits fluorescence, a structure that emits excimer emission, and a structure that emits exciplex emission. It is especially preferable that the chromophore or luminescent phore contains a structure that emits fluorescence. The fluorescence emitted by the luminescent dye molecule makes it easier to analyze with various measuring devices.

[0064] It is preferable that the luminescent dye molecule exhibits multiple types of emission when irradiated with light in the wavelength range of 300 to 400 nm. Because the luminescent dye molecule exhibits multiple types of emission, a special light source is not required when analyzing a subject, and the analysis is less likely to damage the subject.

[0065] The molecular weight of the luminescent dye molecule is appropriately selected depending on the type of chromophore or luminescent phose, the length of the main chain, etc. The molecular weight is preferably in the range of 500 to 10,000, and more preferably in the range of 100 to 4,000. A molecular weight of 10,000 or less results in moderately low specificity to the test subject, allowing the luminescent dye molecule to react non-specifically with multiple locations in the target substance within the test subject.

[0066] Luminescent dye molecules can be synthesized by the following procedure: Prepare monomers by bonding a chromophore or luminescent ester and a phosphate ester to a pentose or hexose. Polymerize these monomers in the desired sequence using the phosphoramidide method with a DNA / RNA synthesizer or the like.

[0067] In this synthesis method, as shown in Figure 7, for example, multiple types of monomers with different chromophores or luminescent phoses are prepared, and the desired number of monomers can be combined by changing the sequence of these monomers. In the example shown in Figure 7, three types of monomers with chromophores or luminescent phoses A, B, and C, respectively, are prepared, and the three monomers can be combined by changing the sequence of these monomers.

[0068] In other words, a wide variety of luminescent dye molecules can be synthesized from multiple types of monomers with different chromophores or luminescent phores. In the example shown in Figure 7, 27 different luminescent dye molecules can be synthesized. By changing the type of monomer used or the number of monomer bonds, a very large number of luminescent dye molecules can be synthesized.

[0069] The luminescent dye molecules according to this embodiment have a main chain structure similar to naturally occurring substances such as DNA and RNA, and can easily interact with various target substances contained in the sample. Therefore, by using luminescent dye molecules, data about the target substances can be easily and in detail. Furthermore, the luminescent dye molecules exhibit multiple types of emission when irradiated with light at a specific wavelength. Therefore, complex and large amounts of data can be obtained depending on the state of the target substance.

[0070] The following describes an example of a method for measuring and acquiring signal data using luminescent dye molecules, but the measurement method is not limited to this method.

[0071] Figure 8 shows a flowchart of the measurement method. A microwell plate is prepared for the interaction of the luminescent dye molecule and the test subject, and either the luminescent dye molecule or the test subject is placed in the reaction field (well) (first component placement step: step S101). The placed luminescent dye molecule or test subject is referred to as the first component.

[0072] The method for placing the first component in each reaction site is not particularly limited and can be appropriately selected depending on the type and physical properties of the first component. Examples of methods for placing the first component include coating by inkjet, coating by dispenser, placement of a carrier supporting the first component, and direct fixation of the first component to the reaction site. Among these, the inkjet method is preferred. By using the inkjet method, liquid first components can be efficiently placed in a large number of reaction sites, and a large amount of data can be acquired.

[0073] If the first component is a luminescent dye molecule and the plate has multiple reaction fields, the same type of first component may be placed in all of the reaction fields, or multiple types of first components may be placed in the same reaction field. Alternatively, different types of first components may be placed in two or more reaction fields. By placing different types of first components in different reaction fields, multiple types of interactions occur between the luminescent dye molecule and the target substance, allowing for a more detailed analysis of the target substance.

[0074] The measurement is started, and first signal data is acquired from the plate on which the first component is placed (first signal data acquisition step: step S102).

[0075] A luminescent dye molecule or the subject, as the other component of the first component, is placed in the reaction field of the plate from which the first signal data was acquired (second component placement step: step S103). The placed luminescent dye molecule or subject is referred to as the second component.

[0076] The second signal data is acquired from the plate containing the second component (second signal data acquisition step: step S104), and the measurement is terminated.

[0077] (Communication Unit) The communication unit 12 transmits the data measured by the measurement unit 11 to the communication unit 23 of the analysis device 20. The communication unit 12 may use either wireless communication or wired communication.

[0078] (Control Unit) The control unit 13 is a processor that provides overall control over the operation of the measuring device 10. The control unit 13 includes a CPU (Central Processing Unit) that performs various calculations, and RAM (Random Access Memory) that provides the CPU with a working memory space and stores temporary data.

[0079] [Analysis Device] (Display Unit) The display unit 21 displays various information on the screen based on the display control signal received from the control unit 24. The display unit 21 is equipped with a display device. Examples of display devices include a display, a projector, etc. The display unit 21 displays the results of the analysis performed by the control unit 24 based on the data measured by the measurement unit 11 of the measuring device 10 and notifies the user.

[0080] (Operation Unit) The operation unit 22 accepts various inputs through user operation. The operation unit 22 is equipped with input devices. Examples of input devices include keyboards, mice, various switches, touchscreens, touchpads, etc.

[0081] (Communication Unit) The communication unit 23 receives data transmitted from the communication unit 12 of the measuring device 10. The communication unit 23 may use either wireless communication or wired communication.

[0082] (Control Unit) The control unit 24 is a processor that provides overall control over the operation of the analysis device 20. The control unit 24 includes a CPU (Central Processing Unit) that performs various calculations, and RAM (Random Access Memory) that provides the CPU with working memory space and stores temporary data.

[0083] The control unit 24 acquires first data about the characteristics of the subject. The first data may be data measured by the measuring device 10, or other data. At this time, the control unit 24 functions as a first acquisition unit. The control unit 24 also acquires second data about the characteristics of the subject. At this time, the control unit 24 functions as a second acquisition unit.

[0084] The control unit 24 generates a trained model based on a group of subject data, including first and second data acquired for known subjects. At this time, the control unit 24 functions as a learning unit. The control unit 24 inputs first data acquired for an unknown subject into the generated trained model and outputs the analysis results for the unknown subject. At this time, the control unit 24 functions as an analysis unit. Specifically, the CPU reads the stored program and loads it into RAM, and works in cooperation with the program loaded into RAM to perform data acquisition, learning, and analysis.

[0085] [Subject] The type of subject in this embodiment is not particularly limited and may be a substance whose structure is known or a substance whose structure is unknown. The subject may be a mixture of various compounds, etc. The subject may belong to any field, such as the medical field, the industrial field, or the food field. The measurement unit 11 may measure the entire subject or measure a specific target substance within the subject.

[0086] Examples of target substances belonging to the medical field include proteins, antibodies, antibody-containing beads, and tumor markers. Examples of subjects containing these target substances include sweat and blood. Examples of target substances belonging to the industrial field include metal ions, metal nanoparticles, carbon nanotubes, magnetic fluids, nanosilica, and crystalline zirconia. Examples of subjects containing these target substances include rainwater, rivers, small ponds, tropical fish breeding water, soil, and wastewater. In other words, the analytical device of this embodiment can be applied to water quality management. Examples of target substances belonging to the food field include proteins, lipids, and carbohydrates. Examples of subjects containing these target substances include food products in general.

[0087] [Operation of the Analysis System] Figure 9 is a flowchart showing an example of the operation of the analysis system 100. When the analysis of the subject is started, the following operations are performed.

[0088] First, the control unit 24 instructs the measurement unit 11 of the measuring device 10 to measure the characteristics of a known subject (Step S1: Measurement step of a known subject).

[0089] Here, "known" means that second data about the characteristics of the subject can be obtained, that is, the second data acquisition step S3 can be performed. The second data consists of items from the information about the subject that are subject to analysis, and examples of the second data include the structure and content of compounds contained in the subject. Furthermore, if the subject is food or the like, examples of the second data include sensory evaluations of the subject such as flavor, texture, and smell.

[0090] In this embodiment, multiple types of known subjects are prepared. For example, when analyzing the flavor of food such as milk, multiple types of milk of different kinds are prepared as known subjects. Here, "different types" means that the components or flavors contained in the milk are different. From the viewpoint of prediction accuracy and robustness, it is preferable to have a larger number of types of known subjects.

[0091] Next, the control unit 24 acquires first data about the characteristics of a known subject (Step S2: Acquisition of first data of a known subject). At this time, the control unit 24 functions as a first acquisition unit. The number of types of first data is not particularly limited and may be multiple types. For example, two types of data may be acquired as first data: data about the optical properties whose behavior changes when interacting with the subject, and the pH value of the subject.

[0092] Next, the control unit 24 acquires second data about the characteristics of a known subject (Step S3: Second data acquisition step for a known subject). At this time, the control unit 24 functions as a second acquisition unit. The number of types of second data is not particularly limited and may be multiple types. For example, two types of data may be acquired as second data: the subject's taste preference and its smell preference.

[0093] Next, the control unit 24 generates a trained model based on the subject data group, which includes the first and second data obtained for known subjects (Step S4: Training Process). At this time, the control unit 24 functions as a learning unit. Details of the training process will be described later.

[0094] Next, the control unit 24 instructs the measurement unit 11 of the measuring device 10 to measure the characteristics of the unknown subject (Step S5: Measurement step of the unknown subject).

[0095] Next, the control unit 24 acquires first data about the characteristics of an unknown subject (Step S6: First data acquisition step for an unknown subject). At this time, the control unit 24 functions as a first acquisition unit.

[0096] Next, the control unit 24 inputs the first data acquired for the unknown subject into the generated trained model, outputs the analysis results for the unknown subject (step S7: analysis process), and terminates the operation. Note that the items output as "analysis results" here include the items of the second data. For example, if the items of the second data are sensory evaluations, the items output as analysis results will include sensory evaluations.

[0097] The details of the learning process are described below. (Learning Process) In the learning process S4, training data is generated from multiple types of subject data sets, and a trained model is generated from the training data. Here, "subject data set" refers to the collection of first data acquired by the first acquisition unit and second data acquired by the second acquisition unit for subjects of the same type.

[0098] For training data, explanatory variables can include numerical values ​​representing the characteristics of the first data set, and numerical values ​​calculated from them. If the first data set is a spectral distribution, explanatory variables can include the intensity of light at each wavelength. The second data set is used as the target variable.

[0099] Figure 10 is a flowchart showing an example of the operation of the learning process S4 in the analysis system 100. When the learning process S4 starts, the following operations are performed.

[0100] First, the control unit 24 determines whether to remove outliers included in the first data acquired by the first acquisition unit (step S41: first data outlier removal determination step). If the control unit 24 determines to remove the outliers (step S41: YES), it removes the outliers (step S42: first data outlier removal step). If the control unit 24 determines not to remove the outliers (step S41: NO), it proceeds to step S43.

[0101] The method for removing outliers is not particularly limited, and known methods can be used. Removing outliers can improve prediction accuracy. Examples of methods for removing outliers include methods for identifying outliers based on quantiles, methods for identifying outliers based on robust estimation (Hubue's method, Cauchy's method, quartile method, principal component analysis, etc.), K-nearest neighbors (K-Means cluster analysis), hierarchical cluster analysis, normal mixture analysis, latent class analysis, and variable clustering. Outliers can also be removed by standardizing the data. Among these, Cauchy's method is preferred from the viewpoint of robustness.

[0102] Next, the control unit 24 tests for significant differences between the subject data groups using a statistical hypothesis testing method (step S43: significance test step). Hereinafter, the "statistical hypothesis testing method" will also be simply referred to as "testing".

[0103] For the second dataset as well, outliers may be removed before testing, just as with the first dataset. The method described above can be used to remove outliers.

[0104] When there are two known types of subjects, for example, the following testing methods can be used. If the data sets for each subject are normal, a parametric test such as the t-test can be used. If the data sets for each subject are not normal, nonparametric tests such as the Mann-Whitney U test and the sign test can be used. However, the testing methods are not limited to these.

[0105] If there are three or more known types of subjects, first, test for significant differences between data groups of multiple subject types. Then, test (post-hoc test) to determine which data groups of subjects have significant differences.

[0106] For testing for statistical significance in the first stage, the following testing methods can be used, for example. If the data set for each subject is normal, the testing method can be a parametric test, such as one-way analysis of variance or two-way analysis of variance. Use one-way analysis of variance if there is one factor that changes the data value, and use two-way analysis of variance if there are two factors. If the data set for each subject is not normal, the testing method can be a nonparametric test, such as the Kruskal-Wallis test or the Friedman test.

[0107] For post-hoc testing in the second stage, multiple comparison methods can be used. While there are no particular limitations on the multiple comparison method, Tukey's method is preferable from the viewpoint of high accuracy of the test. However, Tukey's method can only be applied if the variances are equal and the data set of each subject is normal. If the number of data points differs in each data set of subjects, Tukey-Kramer's method may be used. Other multiple comparison methods, such as Bonferroni's method, Dunnett's method, and Newman-Coyles' method, may also be used depending on the conditions.

[0108] In statistical testing, a significance level (p-value) may be set. The "significance level" refers to the probability of misinterpreting a non-significant difference as a significant difference. For example, setting p = 0.05 means there is a 5% probability of misinterpreting a non-significant difference as a significant difference. Therefore, from the perspective of reducing misinterpretations, a relatively small significance level p-value, such as p = 0.01, may be set. On the other hand, from the perspective of being able to extract data that may have a significant difference even if there are misinterpretations, a relatively large significance level p-value, such as p = 0.1, may be set.

[0109] Between data groups of the subjects, if the p-value at the significance level is below the set value, it can be determined that there is a statistically significant difference; if it is above the set value, it can be determined that there is no statistically significant difference.

[0110] Next, the control unit 24 selects combinations of subject data sets that have been determined to have a significant difference and generates training data (step S44: training data generation step).

[0111] The subject data sets are selected so that combinations of subject data sets that were determined to have a statistically significant difference based on the above test are included in the training data. In other words, the subject data sets are selected so that combinations of subject data sets that were determined not to have a statistically significant difference are not included in the training data.

[0112] For example, by setting the significance level p-value to a relatively small value, the criteria for testing for statistical significance are made stricter, so that only the highly statistically significant group of subject data is selected as training data, thereby improving prediction accuracy.

[0113] Generally, devices that measure the behavior of light have good sensitivity and can acquire complex data, but on the other hand, the data is easily altered by even slight factors. Therefore, variations in the acquired data are likely to occur due to individual differences that occur during the manufacturing of the device, minute changes in the light-emitting probes contained within the device, etc.

[0114] If the first data is, for example, emission spectrum data measured using the above-mentioned emission probe, then even if the same sample of the subject is measured multiple times, differences are likely to occur in the data obtained each time. If the trained model determines that this difference is due to differences in the subject itself, then even if the same sample of the subject is analyzed, the analysis results will change each time. However, for the analysis device 20, it is preferable that the analysis results match even when the same sample of the subject is measured multiple times. In this specification, the degree to which the analysis results match when the same sample of the subject is measured multiple times is referred to as "robustness (reproducibility)". A higher degree of agreement in the analysis results indicates higher robustness.

[0115] From the perspective of improving prediction accuracy, it is preferable to set strict criteria for statistical significance and select only highly significant sample data sets for training. On the other hand, from the perspective of robustness, it is preferable not to set the criteria for statistical significance too strictly and select a reasonable number of sample data sets for training.

[0116] Table I below shows an example of the relationship between training data selection rate, prediction accuracy, and robustness. Here, "training data selection rate" refers to the ratio of the number of subject data sets selected for training after testing, relative to the total number of subject data sets obtained for known subjects. Note that training data selection rates of 90%, 80%, and 60% mean that subject data sets with low significance were excluded at a rate of 10%, 20%, and 40%, respectively.

[0117]

[0118] Examples 1 and 2 show that setting stricter significance criteria and reducing the selectivity of training data improves both prediction accuracy and robustness. Furthermore, Examples 2 and 3 show that setting even stricter significance criteria and further reducing the selectivity of training data improves prediction accuracy but decreases robustness. From this, it can be seen that high prediction accuracy and robustness can be obtained by not making the selectivity of training data too low.

[0119] As an example, we have described the case where the first data is emission spectral data measured using the above-mentioned emission probe, but the type of first data is not limited to this. If the first data is data that is prone to variations with each measurement, high robustness can be obtained by not setting the selectivity of the training data too low, as described above.

[0120] The selectivity of training data that yields high prediction accuracy and robustness varies depending on the type of prediction algorithm used in machine learning. Therefore, the selectivity of training data may be adjusted according to the type of prediction algorithm used in machine learning.

[0121] Next, the control unit 24 generates a trained model using machine learning with the generated training data (step S45: trained model generation step), and terminates the operation of the training step S4.

[0122] Machine learning may be reinforcement learning, deep learning, or deep reinforcement learning. "Reinforcement learning" refers to a learning method that learns the "optimal sequence of actions" through trial and error. "Deep learning" refers to a learning method that learns the features contained in a large amount of data in a stepwise, deeper way. "Deep reinforcement learning" refers to a learning method that combines reinforcement learning and deep learning.

[0123] Analysis in machine learning can be performed using the statistical analysis software "JMP16.2" or "JMPpro16.2" manufactured by SAS Institute Japan Co., Ltd.

[0124] Examples of linear regression prediction algorithms used in machine learning include principal component analysis (PCA), cluster analysis (hierarchical clustering, k-means clustering, and normal mixture clustering), linear discriminant analysis (LDA), and partial least squares regression (PLS regression). These algorithms may also be used in combination.

[0125] Other nonlinear regression prediction algorithms used in machine learning include decision trees, random forests, bootstrap forests, neural networks, K-nearest neighbors, simple Bayes, support vector machines (SVM), nominal logistic regression (multiple logits), and generalized regression (Ridge, Lasso).

[0126] In particular, when the second data set tends to show significant variability between data points, especially when the second data set is a sensory evaluation, the first and second data sets often exhibit a nonlinear relationship. Therefore, from the standpoint of prediction accuracy, it is preferable for the prediction algorithm to be a nonlinear regression.

[0127] The operation of the analysis system 100 will be specifically explained below, using the analysis of milk flavor as an example.

[0128] First, sixteen types of milk (Milk [1] to Milk

[16] ) are prepared as known test subjects. These are all different types of commercially available products, and their ingredients and flavors differ.

[0129] The first data set is the emission spectrum data measured by the measuring device 10 using the above-mentioned emission probe. The emission spectrum data is measured five times for each subject. The control unit 24 of the analysis device 20 acquires the measurement results.

[0130] The second set of data consists of a five-point rating of taste preference (like or dislike) by panelists, with each sample being evaluated by 10 panelists. The user of the analysis system 100 inputs the evaluation results of the 10 panelists (10 patterns) from the operation unit 22 of the analysis device 20. The control unit 24 of the analysis device 20 acquires the results of the sensory evaluation. As a result, the sample data set consists of emission spectrum data (first data) and taste preference evaluation data (second data) for each milk product.

[0131] Next, we determine whether to remove outliers from the first data set for each milk sample. If we decide to remove them, we remove the outliers.

[0132] Next, we will test for significant differences between the acquired milk data groups using statistical hypothesis testing methods. There are 16 milk data groups. Since there are at least 3, we will first test for significant differences between the milk data groups using one-way ANOVA. After that, we will test which milk data groups have significant differences using the Tukey-Kramer method (post-hoc test).

[0133] Next, based on the test results, combinations of subject data sets that are judged to have a statistically significant difference are selected to generate training data. If a statistically significant difference is found between the data set of milk [1] and the data set of milk [2], all data included in the data sets of milk [1] and milk [2] are used as training data. On the other hand, if no statistically significant difference is found between the data set of milk

[14] and the data set of milk

[15] , all data included in the data sets of milk

[14] and milk

[15] are not used as training data. Similarly, if there is no statistically significant difference between the data set of milk

[14] and the data set of milk

[16] , and no statistically significant difference between the data set of milk

[15] and the data set of milk

[16] , then statistically significant differences are found for other milk combinations. In this case, the data sets of milk [1] to

[13] are used as training data, and the data sets of milk

[14] to

[16] are not used as training data. The training data selection rate in this case is 81%.

[0134] Next, a trained model is generated using machine learning with the generated training data. A neural network, which is a nonlinear regression algorithm, is used as the prediction algorithm.

[0135] Next, prepare the milk to be analyzed (milk [X]). The first data will be the emission spectrum data measured by the measuring device 10 using the emission probe described above. The emission spectrum data will be measured five times for each sample.

[0136] The generated pre-trained model is input with the first data set for milk [X], and the analysis results are output. The analysis results output a rating score on a 5-point scale for taste preference (like or dislike). Note that outliers may be removed from the first data set for milk [X] before inputting it into the pre-trained model. The method described above can be used to remove outliers.

[0137] Generally, performing sensory evaluations with high accuracy requires evaluations by many panelists. However, the analysis system 100 of this embodiment does not require evaluations by many panelists, thus simplifying the sensory evaluation process and enabling highly accurate sensory evaluations.

[0138] In this embodiment, the analysis device 20 comprises a first acquisition unit (control unit 24), a second acquisition unit (control unit 24), a learning unit (control unit 24), and an analysis unit (control unit 24). The first acquisition unit acquires first data about the characteristics of the subject. The second acquisition unit acquires second data about the characteristics of the subject. The learning unit generates a trained model based on a group of subject data, including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for known subjects. The analysis unit inputs the first data acquired by the first acquisition unit for an unknown subject into the trained model generated by the learning unit and outputs the analysis results for the unknown subject. The learning unit tests for significant differences between groups of subject data using statistical hypothesis testing methods, selects combinations of subject data groups that are determined to have significant differences, and generates training data. This improves the predictive accuracy of the analysis when using training data with large variability among the data.

[0139] In this embodiment, it is preferable that the second data includes data on the sensory evaluation of the subject. This improves the predictive accuracy of the analysis, especially in sensory evaluation, which is generally difficult to analyze.

[0140] In this embodiment, the statistical hypothesis testing method is preferably a multiple comparison method. This allows for a number of known sample types in the test to be three or more.

[0141] In this embodiment, the multiple comparison method is preferably the Tukey-Kramer method. This allows for high accuracy of the test and enables testing even when the number of data points differs in each subject's data set.

[0142] In this embodiment, it is preferable that the first data includes data on optical properties whose behavior changes in interaction with the subject. This allows for obtaining a large amount of complex data about the subject.

[0143] In this embodiment, it is preferable that the first data includes data on the optical properties derived from a light-emitting probe whose luminescence behavior changes upon interaction with the subject. This allows for the acquisition of accurate data on the target substance contained in the subject, thereby improving the predictive accuracy of the analysis.

[0144] In this embodiment, the first data includes data on the optical properties of the organic electroluminescent element. The organic electroluminescent element preferably has a light-emitting layer 122 containing a light-emitting probe between the electrodes (between the transparent electrode 112 and the counter electrode layer 115). This allows for the acquisition of various data on the target substance contained in the sample, improving the predictive accuracy of the analysis.

[0145] In this embodiment, it is preferable that the luminescent probe has a main chain having one or more structural units including a sugar structure derived from a pentose or hexose and a phosphate ester bond attached to the sugar structure, and one or more chromophores or luminescent phosphodiphores attached to the sugar structure. This allows the luminescent probe to easily interact with target substances having a main chain structure similar to that of DNA, RNA, etc. As a result, data about the target substance can be easily and in detail, and the predictive accuracy of the analysis can be improved.

[0146] In this embodiment, it is preferable that the learning unit removes outliers from the first data acquired by the first acquisition unit to generate training data. This improves the prediction accuracy of the analysis.

[0147] In this embodiment, the method for removing outliers is preferably the Couchy method. This improves the robustness of the analysis.

[0148] In this embodiment, the prediction algorithm used to generate the trained model is preferably a nonlinear regression algorithm. This improves the prediction accuracy of the analysis.

[0149] In this embodiment, the analysis system 100 comprises an analysis device 20 and a measuring device 10 for measuring first data. The first acquisition unit acquires the first data measured by the measuring device 10. This improves the prediction accuracy of the analysis.

[0150] In this embodiment, the analysis method executed by the analysis device 20 includes a first acquisition step (first data acquisition step S2 for known subjects and first data acquisition step S6 for unknown subjects), a second acquisition step (second data acquisition step S3), a learning step S4, and an analysis step S7. The first acquisition step acquires first data about the characteristics of the subjects. The second acquisition step acquires second data about the characteristics of the subjects. The learning step S4 generates a trained model based on a group of subject data including the first data acquired in the first acquisition step and the second data acquired in the second acquisition step for known subjects. The analysis step S7 inputs the first data acquired in the first acquisition step for unknown subjects into the trained model generated by the learning step S4 and outputs the analysis results for the unknown subjects. The learning step S4 tests the significant differences between the subject data groups using statistical hypothesis testing methods, selects combinations of subject data groups that are determined to have significant differences, and generates training data. This improves the prediction accuracy of the analysis.

[0151] In this embodiment, the program causes the computer of the analysis device 20 to function as a first acquisition unit (control unit 24), a second acquisition unit (control unit 24), a learning unit (control unit 24), and an analysis unit (control unit 24). The first acquisition unit acquires first data about the characteristics of the subject. The second acquisition unit acquires second data about the characteristics of the subject. The learning unit generates a trained model based on a group of subject data, including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for known subjects. The analysis unit inputs the first data acquired by the first acquisition unit for unknown subjects into the trained model generated by the learning unit and outputs the analysis results for the unknown subjects. The learning unit tests for significant differences between groups of subject data using statistical hypothesis testing methods, selects combinations of subject data groups that are determined to have significant differences, and generates training data. This improves the prediction accuracy of the analysis.

[0152] This disclosure makes it possible to improve the predictive accuracy of analysis when using training data that has large variability among the data.

[0153] 10 Measuring device 11 Measuring unit 12 Communication unit 13 Control unit 20 Analysis device 21 Display unit 22 Operation unit 23 Communication unit 24 Control unit 100 Analysis system 111 Transparent substrate 112 Transparent electrode 113 Hole transport layer 114 Receptor layer 115 Counter electrode layer 121 Subject layer 122 Light-emitting layer

Claims

1. An analysis device comprising: a first acquisition unit for acquiring first data about the characteristics of a subject; a second acquisition unit for acquiring second data about the characteristics of the subject; a learning unit for generating a trained model based on a group of subject data including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for a known subject; and an analysis unit for inputting the first data acquired by the first acquisition unit for an unknown subject into the trained model generated by the learning unit and outputting analysis results for the unknown subject, wherein the learning unit tests for significant differences between the groups of subject data using statistical hypothesis testing methods, selects combinations of the groups of subject data that are determined to have significant differences, and generates training data.

2. The analytical apparatus according to claim 1, wherein the second data includes data on the sensory evaluation of the subject.

3. The analytical apparatus according to claim 1, wherein the statistical hypothesis testing method is a multiple comparison method.

4. The analysis apparatus according to claim 3, wherein the multiple comparison method is the Tukey-Kramer method.

5. The analytical apparatus according to claim 1, wherein the first data includes data on optical properties whose behavior changes in interaction with the subject.

6. The analytical apparatus according to claim 5, wherein the first data includes data on optical properties derived from a light-emitting probe whose light emission behavior changes upon interaction with the subject.

7. The analysis apparatus according to claim 6, wherein the first data includes data on the optical properties of an organic electroluminescent element, and the organic electroluminescent element has a light-emitting layer including the light-emitting probe between electrodes.

8. The analytical apparatus according to claim 6, wherein the light-emitting probe comprises a main chain having one or more structural units including a sugar structure derived from a pentose or hexose and a phosphate ester bond attached to the sugar structure, and one or more chromophores or light-emitting phosphodiphores attached to the sugar structure.

9. The analysis apparatus according to claim 1, wherein the learning unit generates the learning data by removing outliers included in the first data acquired by the first acquisition unit.

10. The analysis apparatus according to claim 9, wherein the method for removing outliers is the Couchy method.

11. The analysis apparatus according to claim 1, wherein the prediction algorithm used to generate the trained model is nonlinear regression.

12. An analysis system comprising: an analysis device according to claim 1; and a measuring device for measuring the first data, wherein the first acquisition unit acquires the first data measured by the measuring device.

13. An analysis method performed by an analysis device, comprising: a first acquisition step of acquiring first data about the characteristics of a subject; a second acquisition step of acquiring second data about the characteristics of the subject; a learning step of generating a trained model based on a group of subject data including the first data acquired in the first acquisition step and the second data acquired in the second acquisition step for a known subject; and an analysis step of inputting the first data acquired in the first acquisition step for an unknown subject into the trained model generated by the learning step and outputting an analysis result for the unknown subject, wherein the learning step tests for significant differences between the groups of subject data using a statistical hypothesis testing method, selects combinations of the groups of subject data that are determined to have significant differences and generates training data.

14. A program that causes the computer of an analysis device to function as an analysis unit comprising: a first acquisition unit that acquires first data about the characteristics of a subject; a second acquisition unit that acquires second data about the characteristics of the subject; a learning unit that generates a trained model based on a group of subject data including the first data acquired by the first acquisition unit and the second data acquired by the second acquisition unit for a known subject; and an analysis unit that inputs the first data acquired by the first acquisition unit for an unknown subject into the trained model generated by the learning unit and outputs analysis results for the unknown subject, wherein the learning unit is a program that tests for significant differences between the groups of subject data using statistical hypothesis testing methods, selects combinations of the groups of subject data that are determined to have significant differences, and generates training data.

Images