A method and program for evaluating the reliability of observed values.

Abstract

The present disclosure provides a method for evaluating a likelihood that a subject belongs to a certain group with respect to a classification attribute which has a binary classification, the method comprising: receiving a subject score regarding an observation value for the subject; using a relational expression established between the sensitivity and specificity with a score regarding the observation value serving as a parameter to acquire the sensitivity and specificity of the subject score when the subject score serves as the parameter; acquiring a prior probability for the attribute of the subject; and acquiring the likelihood that the subject belongs to the classification attribute peculiar to the subject on the basis of the sensitivity, specificity, and prior probability of the subject.

Description

Technical Field

[0001] The present disclosure relates to a method, program, and system for evaluating the certainty of observed values.

Background Art

[0002] So far, inspection results have been judged as to whether the subject is positive or negative in light of the ROC curve. This has used a specific point on the ROC curve (the point closest to the upper left corner, the point defined by the Youden Index) as the cut-off point. Generally, if the observed value is higher than the cut-off point, a positive result is returned, and if it is lower, a negative result is returned. Based on that cut-off point, the positive predictive value (PPV) has been calculated. However, this positive predictive value is a value specific to the ROC curve and is merely a value for evaluating the evaluation system.

Summary of the Invention

[0003] There is a desire for a method for a subject to more appropriately evaluate whether they are positive or negative from the inspection results, that is, the subject's result.

[0004] According to one embodiment of the present disclosure, a novel method for evaluating the likelihood (certainty) that a subject belongs to a certain group is provided. In some embodiments, the method may include receiving a subject score regarding the observed value of the subject. In some embodiments, the method uses a relational expression established between sensitivity and specificity with the score regarding the observed value as a mediating variable, and when the subject score is used as the mediating variable, the sensitivity of the subject score and specificity and may include obtaining the same. In some embodiments, the method may include obtaining the prior probability of the subject's attribute. In some embodiments, the method may include obtaining the likelihood of belonging to a classification attribute specific to the subject based on the sensitivity, specificity, and prior probability of the subject.

[0005] Further aspects and advantages of the present disclosure will be readily apparent to those skilled in the art from the following detailed description, which shows and describes only exemplary embodiments of the present disclosure. As will be understood, other different embodiments are possible, and some of their details can be modified in various obvious ways without departing from the present disclosure. Accordingly, the drawings and description should be considered illustrative and not limiting in nature. [Brief explanation of the drawing]

[0006] [Figure 1] A flowchart illustrating an evaluation method according to one embodiment of this disclosure is shown. [Figure 2] A block diagram showing a computer control system according to one embodiment of this disclosure is shown. [Modes for carrying out the invention]

[0007] Figure 1 shows a flowchart of a novel method for evaluating the likelihood (probability) of an object belonging to a certain group, according to one embodiment of the present disclosure. In step S101, the object score for the observed value of the object is received. In step S102, the observed value is Score Using the relationship between sensitivity and specificity with the parameter, we can determine the sensitivity of the target score and the specificity when the target score is used as the parameter. specificity In step S103, the prior probability of the target attribute is obtained. In step S104, the likelihood of belonging to a classification attribute specific to the target is obtained based on the sensitivity, specificity and prior probability of the target.

[0008] <Observation> As used in this disclosure, "observation" is not limited to "the act of observing" in the narrow sense, but generally refers to observation, measurement, and analysis in the fields of biology, medicine, pharmacology, biochemistry, physics, chemistry, electricity, and optics.

[0009] <Test> In some embodiments, the observation may be a clinical test. Clinical testing includes specimen testing, biopsy, imaging, pathological diagnosis, physical examination, psychological testing, and other tests aimed at determining the presence or absence of disease or diagnosing it, or even if not for that purpose, obtaining related information.

[0010] Specimen testing includes biochemical tests, hematological tests, urine and fecal tests, immunological tests, microbiological tests, etc.

[0011] In this disclosure, the body fluid used in the examination means the body fluid obtained from the subject or a sample derived from said body fluid. The body fluid may not be limited to blood, serum, plasma, lymph, tissue fluid such as interstitial fluid, intercellular fluid, or interstitial fluid, or body cavity fluid, serosal fluid, pleural fluid, ascites, pericardial fluid, cerebrospinal fluid, synovial fluid, or aqueous humor. The body fluid may also be digestive fluids such as saliva, gastric juice, bile, pancreatic juice, or intestinal juice, or sweat, tears, nasal mucus, urine, semen, vaginal fluid, amniotic fluid, or breast milk. The body fluid may be animal body fluid or human body fluid.

[0012] Biopsies include respiratory and circulatory function tests, ultrasound examinations, various tests using monitoring devices, electroencephalograms, neuromuscular examinations, otolaryngological examinations, ophthalmic examinations, dermatological examinations, clinical and neuropsychological examinations, stress tests, radioisotope tests, endoscopic examinations, etc. In some embodiments, the biopsy may be a liquid biopsy.

[0013] In some embodiments, the observed values ​​may be quantities, frequencies, or other test values ​​related to gene expression in a genetic test.

[0014] A gene may be nucleic acid (at least one of DNA and RNA). RNA may be messenger RNA (mRNA), transfer RNA (tRNA), Ribosomes RNA (rRNA), microRNA (miRNA), etc., may also be used.

[0015] In genetic testing, a sample of the subject's bodily fluids (blood, saliva, urine, etc.) may be obtained, and a predetermined or arbitrary amount of nucleic acid (relative or absolute) may be measured. The nucleic acid may be amplified. The nucleic acid may also be measured using genetic analysis equipment such as DNA chips (also called microarrays) or sequencers.

[0016] Genetic testing may include testing for gene mutations. Gene mutations may include the measurement of copy number variations. Changes in the number and expression levels of single nucleotide polymorphisms (SNPs) may be measured. Fusion genes may be measured. For example, it may be determined whether or not fusion has occurred at a specific gene or base site. The number of fusion genes may be measured. Chromosomal abnormalities may be measured. The presence or absence of chromosomal abnormalities, their amount or frequency within a specific region, etc., may be measured. Chromosomal abnormalities may be structural changes, changes in chromosome number, or both. Tumor mutational load (TMB) may be measured. The number of tumor genes or the TMG score may be measured. The amount of epigenetic changes such as methylation (number of sites, frequency at specific sites) and acetylation may be measured. The number of sites where these mutations occur or the amount of change at specific sites may be measured. Microsatellite instability (MSI) analysis or testing may be performed. The number of changed bases or their frequency in microsatellite regions may be measured. Splicing abnormalities may be measured. The presence or absence of such abnormalities may be measured, and the number (number of locations or base pairs), absolute number, frequency, etc., of these locations may also be measured.

[0017] If the observation is evaluated numerically, the score may be the observed value itself, or a value converted from that observed value. If the observation does not provide a numerical value, the score may be obtained by quantifying the observation result according to the evaluation method.

[0018] The score may be the value of the test result itself or a processed value. The score may be a normalized value. The score may be calculated based on the value of the test result or an evaluation. The score may be calculated based on the results of multiple tests. In some embodiments, they may be combined and software such as machine learning may be used to calculate the score. The score may be continuous or discontinuous (discrete numbers, e.g., binary 0 / 1). In some embodiments, in a genetic test, the amount of a predetermined gene (e.g., RNA) in a body fluid may be used as the score, or a score obtained from the gene expression level profile may be utilized.

[0019] In some embodiments, the method of the present disclosure may include providing a binary classification based on a score related to an observed value. In some embodiments, the method of the present disclosure may include providing a multi-value classification based on a score related to an observed value.

[0020] The binary (or multi-value) classification may be prepared or provided in advance or independently of the observation of the subject. A predetermined classification may be obtained for the score of the observed value.

[0021] <Relational expression> In some embodiments, a relational expression using the score as a mediating variable may be used. The mediating variable may also be called a threshold or a cut-off point.

[0022] In some embodiments, a relational expression between the value related to one side of the binary classification and the value related to the other side may be used. In some aspects, for the binary classification, either one of the TPR (true positive rate) and the FNR ( False negative rate ) and either one of the FPR (false positive rate) and the TNR ( true negative rateA relational expression with any one of them may be used. For example, any one of the relational expressions of the relational expression between TPR and FPR; the relational expression between TPR and TNR; the relational expression between FNR and FPR; and the relational expression between FNR and TNR may be used. For example, a relational expression between sensitivity (TPR) and specificity (1 - FPR) may be used. For example, a relational expression between sensitivity (TPR) and (1 - specificity) = FPR may be used.

[0023] In some embodiments, the relational expression may be represented by ROC (Receiver Operating Characteristics), or may be represented by an ROC curve.

[0024] Using the score (target score) based on the observed value of the target as the score of the relational expression (that is, target score = score of the relational expression), such as sensitivity and specificity, can be a feature of the present disclosure. Thereby, for example, even if an equation used in Bayesian statistics (Bayesian estimation, Bayesian probability) is used, the evaluation value obtained therefrom is not a value for evaluating the statistical system, but can be used as an evaluation of the observed value of the target.

[0025] <Prior probability> The prior probability in Bayesian statistics may be investigated before (prior to) the observation of the target. In some embodiments, the prior probability may be obtained after the observation of the target, that is, it may be used as the prior probability. In some embodiments, the prior probability may be objectively, for example, after the observation regarding the target person. acquired This may be the case. Using the prior probability obtained after the observation, the likelihood may be calculated, or the previously obtained observation results may be recalculated using a new prior probability.

[0026] In some embodiments, the likelihood that the observed value is positive may be expressed as the conditional probability of the positive predictive value.

[0027] <Likelihood> In some real-world scenarios, the probability (likelihood) that a subject actually belongs to a certain group (class) may be calculated using Bayesian statistics. The likelihood may be calculated using a Bayesian formula for conditional probability, which includes the prior probability.

[0028] For example, by clinical testing The target person To assess whether or not someone has a disease, calculation The expression can be expressed as follows: [Mathematics 1] JPEG0007872037000001.jpg1394 Therefore, the above formula can be rewritten as follows: [Math 2] JPEG0007872037000002.jpg1492 Here, Se, Sp, and α represent sensitivity, specificity, and prevalence, respectively.

[0029] Similarly, the negative predictive value can be expressed as follows:

number

number

[0030] However, the sensitivity and specificity shown in this disclosure are not values ​​inherent to the ROC curve (such as the point closest to the left corner on the ROC curve, or the Youden index). They are sensitivity (target sensitivity) and specificity (target specificity) calculated from the values ​​on the ROC curve corresponding to the scores of the observed values ​​of the subject. Therefore, since the above formulas differ from their classical meanings, they can be called modP (having the disease | positive), modPPV, modN (not having the disease | negative), and modNPV, respectively. The names are not limited to these and may be other.

[0031] In some embodiments, the likelihood may be expressed by a single function or by multiple functions. The sum of multiple functions may be called a single function. For example, multiple functions may be used in combination. For example, multiple functions may be defined by a range of scores. For example, PPV and NPV may be adapted depending on the range of scores. For example, PPV may be used for scores above a certain value, and NPV may be used for scores below that value. In some embodiments, likelihood ratios such as the positive likelihood ratio and the negative likelihood ratio may be used.

[0032] <Multi-value classification> According to some embodiments of this disclosure, the likelihood of an object belonging to each class defined in a multi-class classification may also be evaluated. In some embodiments, a one-versus-one model may be used. In some embodiments, a one-versus-rest model may be used.

[0033] The probability that class i is positive in an N-value classification using a one-to-other model can be expressed as follows:

number

[0034] Alternatively, the value obtained by substituting the target score (mod PPV) can be used as the likelihood.

number

[0035] Table 1 shows a cross-tabulation table for the case where N=3. [Table 1] For example, the likelihood that a test result is true (i.e., "true class 1") can be expressed as follows, based on a one-to-other model and Bayesian statistics:

number

number

[0036] Extending the above, in the case of N-class classification (N classes), the likelihood that the test result is true to be class i (true class i accuracy) can be expressed as follows:

number

number

[0037] <Examples> Using one embodiment of this disclosure, we evaluated whether a subject had cancer based on RNA expression. The results are described below. [Table 2]

[0038] Table 2 shows the scores obtained by a certain lung cancer biomarker testing method and their corresponding evaluation values ​​for each subject (A-D). ROC curves between lung cancer test results and test results (scores) were obtained in advance.

[0039] Subjects A and B share the same attribute: they are both male in their 20s. Subjects C and D share the same attribute: they are both male in their 50s. The prevalence rate is determined by the subject's attribute. The prevalence rate for men in their 20s is 0.00055%. The prevalence rate for men in their 50s is 0.06260%.

[0040] According to the test results, among the men in their 20s, Subject A had a relatively high score (0.62), while Subject B had a relatively low score (0.02). Among the men in their 50s, Subject C had a relatively high score (0.54), while Subject D had a relatively low score (0.12).

[0041] Here, the score is calculated to show zero "0" when it corresponds to the "threshold" or "cutoff point" that separates positive from negative results. Therefore, a score close to zero means that the value is close to the threshold (subjects B and D). On the other hand, a score far from zero means that the value is far from the threshold. Alternatively, a score far from zero in the positive direction suggests a high probability of being positive (subjects A and C).

[0042] The ROC curve for this lung cancer test result shows the sensitivity and specificity based on the Youden index. These sensitivities and specificities depend on the index itself. In Table 2, the sensitivity and specificity corresponding to the Youden index happened to be the same. However, generally speaking, sensitivity and specificity do not need to have the same value.

[0043] The conventional positive predictive value (PPV) is calculated based on the prevalence and sensitivity of these values, which are derived from the Youden index. As a result, subjects A and B, who belong to the same attribute of being male in their 20s, have the same prevalence and therefore have the same PPV. In other words, subjects A and B have the same PPV (0.0039%) despite having different scores. Similarly, subjects C and D, who belong to the same attribute of being male in their 50s, have the same prevalence. Therefore, despite having different scores, they have the same PPV (0.4466%).

[0044] Using the conventional PPV method, if the score is greater than the threshold, all subjects (male subjects A and B in their 20s, or male subjects C and D in their 50s) will be judged as positive. This result does not depend on the magnitude of the test score. Furthermore, using the conventional PPV calculation formula in this case will yield the same PPV value in all cases. This is because PPV is a parameter that represents the characteristics of the ROC curve. The conventional method could not represent differences in test scores. The conventional method could not individually express the likelihood of each subject having the disease.

[0045] On the other hand, the evaluation values ​​according to one embodiment of this disclosure use the sensitivity and specificity when the score is used as a parameter in the ROC curve. Subjects A to D each have different scores, and therefore their sensitivity and specificity are also different (Table 2). The evaluation values ​​were obtained by substituting these sensitivity and specificity, along with the prevalence, into a general PPV calculation formula. Therefore, it differs from conventional PPV.

[0046] As shown in Table 2, subjects A through D were given different evaluation values ​​(0.3808%, 0.0044%, 18.2769%, and 0.9295%, respectively). Subjects A and B, who have the same attributes, have different evaluation values. Similarly, subjects C and D, who have the same attributes, have different evaluation values. In this way, for multiple subjects with the same attributes but different scores, it is possible to assign evaluation values ​​that reflect the score level and prevalence, representing the likelihood of having the disease. Furthermore, subject D obtained a higher score than subject A, but a lower evaluation value. In this way, it is possible to assign comparable evaluation values ​​(e.g., likelihood of having the disease) for multiple subjects belonging to different attributes.

[0047] Thus, it has been shown that the method described in this disclosure can be used to evaluate the likelihood of a subject having cancer more appropriately than in conventional methods.

[0048] <Other application examples> The methods disclosed herein can be applied to clinical testing, but are not limited to that; they can also be applied to other forms of evaluation.

[0049] In some embodiments, the observation may be a medical image. Some embodiments of this disclosure may also be applied to medical image processing (medical image, image-based disease diagnosis).

[0050] Images include, for example, optical images, ultrasound images, X-ray images, magnetic resonance imaging (MRI), and radioisotope (RI) images, without limitation. Medical imaging diagnosis may also refer to general radiographic examinations (commonly known as radiographs). For example, it may include simple radiographic diagnosis or dental panoramic radiographic diagnosis. others, Medical imaging diagnostics may include mammography, computed tomography (CT), gastrointestinal contrast radiography (barium examination), interventional radiology (IVR), etc.

[0051] The score in diagnostic imaging may be determined based on a continuous confidence method. The score may also be determined based on calculation methods or algorithms such as artificial intelligence or software.

[0052] Some embodiments of this disclosure can be applied to various evaluations other than clinical tests or similar tests.

[0053] Some embodiments of this disclosure provide applications for personal authentication. In some embodiments, observation for biometric authentication may be performed. Biometric authentication includes, but is not limited to, fingerprint authentication, knuckle print authentication, vein authentication (on fingers, palms, backs of hands, etc.), palm (palm) authentication, iris authentication, 2D / 3D facial recognition, optical or X-ray dental imaging authentication, voiceprint authentication, and handwriting authentication. Images, sound waveforms, and other data can be used as observed values.

[0054] Some embodiments of this disclosure provide applications to meteorological forecasting. In some embodiments, the likelihood of a certain weather event occurring may be evaluated. For example, based on various meteorological data (pressure patterns, humidity, temperature, wind speed, wind direction, jet stream conditions, sea surface temperature, tidal currents, topography, etc.), the probability of precipitation, solar radiation, wind speed, etc., may be predicted.

[0055] Some embodiments of this disclosure provide applications to natural disaster prediction. In some embodiments, the likelihood of a particular natural disaster occurring may be evaluated. For example, the likelihood of natural disasters such as heavy rain, floods, landslides, wildfires, earthquakes, and volcanic eruptions may be evaluated based on various geological data, meteorological data, radiation data, planetary data, etc.

[0056] Some embodiments of this disclosure provide applications for predicting the occurrence of problems in production lines in industrial fields such as mechanical, electrical, chemical, and pharmaceutical industries. In some embodiments, the likelihood of a particular problem occurring may be evaluated. For example, the likelihood of problems occurring in industrial production, such as machine tools, electrical systems, and chemical reactions, may be evaluated based on various operational data and abnormal signals (such as machine vibration, sound, current, temperature, product characteristics, and changes in equipment status).

[0057] Some embodiments of this disclosure provide applications for forecasting stock prices, national or regional growth rates, inflation rates, and interest rate fluctuations. Some embodiments of this disclosure provide applications for predicting horse racing results. This disclosure is not limited to the applications described herein. This disclosure may have other applications.

[0058] <System> This disclosure also provides a computer-controlled system configured to implement methods provided herein, such as a method for evaluating the likelihood (probability) of an object belonging to a certain group.

[0059] Figure 2 shows one embodiment of a computer system 101 connected to a network 130 for performing the evaluation method of the present disclosure. The computer system 101 shown in Figure 2 is communicably connected to the network 130 and can communicate with a user interface 140 through it. The whole system functions as a network system 100.

[0060] The computer system 101 comprises a central processing unit (CPU, referred to herein as "processor" and "computer processor") 105, memory or memory location 110, electronic storage unit 115, communication interface 120 for communicating with one or more other systems, and peripheral devices 125.

[0061] The CPU 105 may be a single-core or multi-core processor, or multiple processors for parallel processing. The CPU may also be a GPU. The memory 110 may, for example, be random-access memory, read-only memory, or flash memory. The storage unit 115 may be a data storage unit (or data repository) for storing data. The storage unit 115 may, for example, be a hard disk, magnetic tape, etc. The communication interface 120 may, for example, be a network adapter, etc. The communication interface 120 can communicate with the user interface 140 via the network 130. Peripheral devices may, for example, be a cache, other memory, data storage, and / or an electronic display adapter, etc.

[0062] Multiple user interfaces 135 may be connected to the network 130 in a communicative manner. The user interfaces 135 may be located within or connected to the computer system 101.

[0063] In Figure 1, the computer system 101, including the memory 110, storage unit 115, interface 120, and peripheral devices 125, communicates with the CPU 105 via a communication bus (solid line) such as a motherboard.

[0064] One or more components of system 101 may communicate in other ways. One or more components of system 101 may be located in substantially the same place and may be connected in a communicative manner, for example, via network 130.

[0065] The computer system 101 in Figure 1 can be operably coupled to a computer network ("Network") 130 using a communication interface 120. Network 130 may be the Internet, an intranet and / or extranet, or an intranet and / or extranet communicating with the Internet. In some cases, Network 130 is a telecommunications and / or data network. Network 130 may include one or more computer servers that can enable distributed computing, such as cloud computing. In some cases, Network 130 can implement a peer-to-peer network that, with the help of computer system 101, can enable devices coupled to computer system 101 to act as clients or servers.

[0066] The CPU 105 can execute a set of machine-readable instructions that can be implemented by a program or software. Instructions may be stored in a memory location, such as memory 110. Instructions may be directed to the CPU 105, which may then be programmed or otherwise configured to implement the methods of this disclosure. Examples of operations performed by the CPU 105 may include fetching, decoding, executing, and writing back.

[0067] The CPU 105 may be part of a circuit, such as an integrated circuit. One or more other components of system 101 may be included in the circuit. In some cases, the circuit is an application-specific integrated circuit (ASIC).

[0068] The storage unit 115 can store files such as drivers, libraries, and saved programs. The storage unit 115 can also store user data such as user preferences and user programs. In some cases, the computer system 101 may include one or more additional data storage units located outside the computer system 101, such as those located on remote servers that communicate with the computer system 101 via an intranet or the internet. The computer system 101 can communicate with one or more remote computer systems via the network 130.

[0069] The methods described herein can be implemented, for example, by machine-executable code stored in an electronic storage location of a computer system 101, such as memory 110 or an electronic storage unit 115. Alternatively, machine-readable code can be provided in software form. During use, the code can be executed by the processor 105. In some situations, the code can be retrieved from the storage device 115 and stored in memory 110 for immediate access by the processor 105. In some situations, the electronic storage device 115 can be omitted, and the machine-executable instructions are stored in memory 110.

[0070] Code can be pre-compiled and configured for use on a machine with a processor configured to run the code, or it can be compiled at runtime. The code may be provided in a programming language that can be chosen to allow the code to run in either a pre-compiled or compiled form.

[0071] <How the system works> Using a network system 100 that includes such a system 101, the likelihood of an object belonging to a certain group (class) can be evaluated based on the observed results of the object.

[0072] First, observations are made on the target (not shown). Subsequently, observation-related information, such as the observed values ​​and information related to the target, is input to the user interface 135 passively by an operator, automatically from another device connected via communication, or based on commands from an operator.

[0073] Information related to the input observation is entered via the user interface 140, transmitted via the network 130 to the computer system 101, and then received via the communication interface 120. The information received via the communication interface 120 is temporarily stored in memory 110.

[0074] In the embodiment shown in Figure 2, data already obtained in relation to the observation is stored in storage 115. For example, information regarding the attributes of the target. Information Information related to observations such as tests, values ​​and data obtained from observations, scores related to observation values, statistical information such as their raw data and statistical values, classifier algorithms, statistical data after classification, data expressed in binary or multinomial classification, information related to Bayesian statistics such as interclass parameters such as thresholds, sensitivity and specificity, and relationships between them such as ROC curves are stored in storage 115.

[0075] Based on the received information about the target's attributes, the CPU 105 accesses the storage 115 to obtain information related to the target's attributes, such as prior probabilities (prevalence in the case of disease testing), ROC curves, and likelihood evaluation formulas. The obtained information is temporarily stored in the memory 110.

[0076] CPU105 uses the acquired relational expressions, parameters, and information about the subject to calculate or evaluate the likelihood of the subject belonging to a certain group (class).

[0077] The CPU 105 may transmit the results of the likelihood evaluation to a peripheral device 125 for display on a display device, or it may transmit the results to other devices such as a user interface 140 via the network 130 through the communication interface 120.

[0078] The CPU 105 may store information about the subject in the storage 115 based on the results of the likelihood evaluation. The newly stored data in the storage 115 may be incorporated into the population and used in the next evaluation.

[0079] The computer system 101 can be programmed or otherwise configured to adjust one or more parameters to evaluate the likelihood that an object belongs to a certain group (class) based on the observed results of the object.

[0080] Embodiments of systems and methods provided herein, such as computer system 101, can be embodied in programming. Various embodiments of the technology can typically be considered “products” or “manufactured goods” in the form of machine (or processor) executable code and / or associated data held or incorporated in some kind of machine-readable medium. Machine executable code can be stored in memory (e.g., read-only memory, random-access memory, flash memory) or electronic storage units such as hard disks. “Storage” type media can include any or all of the tangible memory of a computer, processor, etc., or their associated modules, such as various semiconductor memories, tape drives, disk drives, etc., which can provide non-temporary storage for software programming at any time. All or part of the software may be communicated in particular through the Internet or various other communication networks. Such communication can enable, for example, the loading of software from one computer or processor to another computer, for example, from a management server or host computer to an application server computer platform. Thus, other types of media that can carry software elements include light, electricity and electromagnetic waves, as used throughout the physical interface between local devices, via wired and optical terrestrial network lines and via various air links. Physical elements that carry such waves, such as wired or wireless links and optical links, can also be considered as media carrying software. As used herein, unless limited to non-temporary, tangible “storage” media, terms such as computer or machine “readable media” refer to any medium involved in providing instructions to a processor for execution.

[0081] Therefore, machine-readable media such as computer executable code can take many forms, including but not limited to tangible storage media, carrier media, or physical transmission media. Non-volatile storage media include optical or magnetic disks, such as any storage device, such as any computer, which may be used to implement a database as shown in the drawings. Volatile storage media include dynamic memory, such as the main memory of a computer platform. Tangible transmission media include coaxial cables, copper wires, optical fibers, and wires that make up buses in computer systems. Carrier transmission media may take the form of electrical or electromagnetic signals, or acoustic or light waves, such as those generated during radio frequency (RF) and infrared (IR) data communications. Therefore, common forms of computer-readable media include, for example, floppy disks, flexible disks, hard disks, magnetic tapes, and other magnetic media; CD-ROMs, DVDs, or DVD-ROMs, and other optical media; punch cards, paper tapes, and other physical storage media with hole patterns; RAM, ROMs, PROMs, and EPROMs, FLASH®-EPROMs, and other memory chips or cartridges; carriers for transmitting data or instructions, cables, or links for transmitting such carriers; or other media from which a computer can read programming code or data. Many of these forms of computer-readable media may be involved in transporting one or more sequences of one or more instructions to a processor for execution.

[0082] The computer system 101 includes, for example, an electronic display 125 having a user interface (UI) 140 for providing signals from a chip over time, or can communicate with such a display. Examples of UIs include, but are not limited to, graphical user interfaces (GUIs) and web-based user interfaces.

[0083] The methods and systems of this disclosure can be implemented by one or more algorithms. The algorithms can be implemented by software at runtime by the central processing unit 105.

[0084] This disclosure provides software for causing a computer or the like to execute the method described herein, and a storage medium for storing such software.

[0085] This disclosure also provides the following embodiments: A101 A method for evaluating the likelihood (probability) of an object belonging to a certain group for classification attributes that have a binary classification, This involves receiving a target score for the target observation, Using the relationship that already exists between sensitivity and specificity with the score relating to the observed value as the parameter, obtain the sensitivity and specificity of the target score when the target score is used as the parameter, Obtaining the prior probability of the aforementioned target attribute, Based on the sensitivity, specificity, and prior probability of the subject, the likelihood of belonging to a classification attribute specific to the subject is obtained. A method for providing this. A102 To obtain the likelihood of belonging to a classification attribute specific to the aforementioned subject. Each of these comprises obtaining the modified positive predictive value or the modified negative predictive value with respect to the aforementioned target score. The method according to Embodiment A101. A201 A method for evaluating the likelihood (probability) that a subject is positive or negative in a clinical test having a binary classification, To receive the target score for the target clinical test, From the relationship between sensitivity and specificity with respect to the score of the clinical test as the mediating variable, the sensitivity and specificity of the target score are obtained when the target score is used as the mediating variable. Obtaining the prevalence rate of the aforementioned target attribute, Based on the sensitivity, specificity, and prevalence of the subject, the likelihood of the subject being positive or negative is obtained. A method for providing this. A202 Obtaining the likelihood that the subject is positive or negative comprises obtaining the modified positive predictive value or modified negative predictive value for the subject score, respectively. The method according to Embodiment A201. A211 The aforementioned clinical test is a biological test. The method according to Embodiment A201 or A202. A212 The aforementioned clinical test is a liquid biopsy. The method according to any one of embodiments A201 to A211. A213 The aforementioned liquid biopsy is a urine test or a blood test. The method according to Embodiment A212. A221 The aforementioned clinical test is a genetic test. The method according to any one of embodiments A201 to A213. A222 The aforementioned genetic test is an RNA test. The method according to Embodiment A221. A223 The aforementioned gene test includes testing for genes derived from urine. The method described in Embodiment A222. A224 The aforementioned gene test comprises examining nucleic acids contained within exosomes. The method according to any one of claims A221 to A223. A225 The exosomes are derived from urine. The method described in Embodiment A224. A301 A method for evaluating the likelihood (probability) of an object belonging to a certain class for a classification attribute with N (where N is a natural number) classifications, This involves receiving a target score for the target observation, In the N-term classification obtained for the score relating to the aforementioned observed value, the true "class i" rate and the false "class i" rate for the aforementioned target score are obtained from the relationship that holds between the probability that class i (1≦i≦N, i is a natural number) is true (true "class i" rate) and the probability that class i is false (false "class i" rate), with the aforementioned score as the parameter. Obtaining the prior probability of the aforementioned target attribute, Based on the true “class i” rate, the false “class i” rate, and the prior probability of the subject, the likelihood of belonging to class i, which is specific to the subject, is obtained. A method for providing this. A302 Based on the true “class i” rate, the false “class i” rate, and the prior probability of the subject, obtaining the likelihood of belonging to class i, which is specific to the subject, comprises obtaining the conditional probability value of the subject score, based on Bayesian statistics. The method according to Embodiment A301. A303 Obtaining the likelihood of belonging to class i, which is specific to the subject, based on the true “class i” rate, the false “class i” rate, and the prior probability of the subject, comprises obtaining the true class i accuracy rate or the false class i accuracy rate with respect to the subject score. The method according to Embodiment A301. B101 A program for causing a computer to perform the method described in any one of embodiments A101 to A303. C101 A computer-readable storage medium for storing the program described in Embodiment B101.

[0086] While preferred embodiments of the present disclosure have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided only as examples. This disclosure is not intended to be limited by any specific examples provided herein. While this disclosure has been described with reference to the preceding specification, the descriptions and illustrations of embodiments herein are not intended to be constrained. Those skilled in the art will be able to conceive of numerous variations, alterations, and substitutions without departing from this disclosure. Furthermore, it should be understood that all aspects of this disclosure are not limited to any specific depictions, configurations, or relative proportions described herein, depending on various conditions and variables. It should be understood that various alternative forms of the embodiments of this disclosure described herein may be used in carrying out the invention. Therefore, this disclosure is intended to also cover such alternatives, alterations, variations, or equivalents. The claims of this application define the scope of the invention, and the methods and structures within these claims, as well as their equivalents, are intended to be covered thereby. [Explanation of symbols]

[0087] 100 Network Systems 101 Computer Systems 105 Central Processing Unit 110 memory 115 Memory Unit 120 Communication Interfaces 125 Peripheral devices 130 Networks 140 User Interfaces

Claims

1. A method for evaluating the likelihood that a subject is positive or negative for a clinical test having a binary classification, To receive the target score for the target clinical test, From the relational expression that holds between sensitivity and specificity with the score related to the clinical test as the mediating variable, the sensitivity and specificity of the target score when the target score is used as the mediating variable are obtained. Obtaining the prevalence rate of the aforementioned target attribute, Based on the sensitivity, specificity, and prevalence of the subject, the likelihood of the subject being positive or negative is obtained. A method for providing this.

2. Obtaining the likelihood that the subject is positive or negative comprises obtaining the modified positive predictive value or modified negative predictive value for the subject score, respectively. The method according to claim 1.

3. The aforementioned clinical test is a biological test. The method according to claim 1 or 2.

4. The aforementioned clinical test is a liquid biopsy. The method according to any one of claims 1 to 3.

5. The aforementioned liquid biopsy is a urine test or a blood test. The method according to claim 4.

6. The aforementioned clinical test is a genetic test. The method according to any one of claims 1 to 5.

7. The aforementioned genetic test is an RNA test. The method according to claim 6.

8. The aforementioned gene test includes testing for genes derived from urine. The method according to claim 7.

9. The aforementioned gene test comprises examining nucleic acids contained within exosomes. The method according to any one of claims 6 to 8.

10. The exosomes are derived from urine. The method according to claim 9.

11. A program for causing a computer to perform the method described in any one of claims 1 to 10.

Images