A memory ability evaluation method based on edit distance

Abstract

The application discloses a memory capacity evaluation method based on an edit distance, comprising the following steps: collecting a subject's original digital string, a subject's answer digital string and interval time; calculating the shortest edit distance of the subject's original digital string and the subject's answer digital string by using an edit distance algorithm; selecting a kernel function, calculating a forgetting information ratio by using the kernel function, combining the interval time to obtain a data point; fitting the data point and a corresponding point on an Ebbinghaus forgetting curve; if the fitting result meets an expected condition, adjusting the parameters of the kernel function according to the fitting result to obtain an optimal kernel function; otherwise, re-fitting / re-selecting the kernel function; calculating the memory rate of the subject by using the optimal kernel function to evaluate the memory capacity; the application is suitable for mixed memory task evaluation, can provide scientific theoretical support, better expand the stimulus type and evaluation mode of the n-back paradigm, and provide a new direction for working memory related research.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of memory evaluation, and in particular to a memory capacity evaluation method based on edit distance. BACKGROUND

[0002] In the working memory related experiment, the demand for memory capacity measurement often appears, for example, in the n-back experiment paradigm, the judgment accuracy of the subject is selected to reflect the memory capacity, and this judgment method is sufficient in the single image or position memory task, but once the experiment involves a mixed memory task, this simple judgment method will lose its scientificity because it is difficult to control the variables. Assuming that in a mixed memory task n-back paradigm, the subject needs to remember a string of random and logical numbers, the subject needs to remember each number and the position of each number, if you want to calculate the average accuracy to represent the memory capacity, you need to ensure that the difficulty of each judgment is the same, then if the memory number string is 123456, is the difficulty of judging 123457 and 123465 the same? If not, which one is more difficult? It can be seen that in the mixed memory task, these problems are difficult to be solved scientifically.

[0003] At present, there is no unified and suitable standard for memory capacity in the experiment involving mixed memory task, and the accuracy evaluation method in the n-back paradigm is mostly used, but this method has been shown to be unsuitable for mixed memory tasks in previous demonstrations, or some standards are designed by referring to the definition of memory span, but the scientificity of these standards is not given. SUMMARY

[0004] This section is intended to summarize some aspects of the embodiments of the present application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of the specification to avoid obscuring the purpose of this section, the abstract and the title, and such simplifications or omissions cannot be used to limit the scope of the present application.

[0005] In view of the above existing problems, the present application is proposed.

[0006] Therefore, the present application provides a memory capacity evaluation method based on edit distance, which solves the problem of poor interpretability of the prior art, and can better expand the stimulus type and evaluation method of the n-back paradigm.

[0007] To solve the above technical problems, the application provides the following technical solutions, comprising: collecting a subject's original digital string, a subject's answered digital string and interval time; calculating the shortest edit distance of the subject's original digital string and the subject's answered digital string by using an edit distance algorithm; selecting a kernel function, calculating a forgetting information ratio by using the kernel function, combining the forgetting information ratio and the interval time to obtain data points; fitting the data points and points on the corresponding Ebbinghaus forgetting curve by using a curve fitting algorithm; if the fitting result meets the expected condition, adjusting the parameters of the kernel function according to the fitting result to obtain the best kernel function; otherwise, re-fitting / re-selecting the kernel function according to the reason for the fitting failure; calculating the memory rate of the subject by using the best kernel function to evaluate the memory ability; wherein the abscissa of the data points is the forgetting information ratio, and the ordinate is the interval time.

[0008] As a preferred scheme of the memory ability evaluation method based on the edit distance, the interval time is selected as the key point of the original memory experiment based on Ebbinghaus.

[0009] As a preferred scheme of the memory ability evaluation method based on the edit distance, when the length of the subject's answered digital string is greater than the length of the subject's original digital string, the original digital string is edited with the digital string of the same length in the subject's answered digital string in sequence, and the minimum edit distance is taken as the final result.

[0010] As a preferred scheme of the memory ability evaluation method based on the edit distance, when the length of the subject's answered digital string is less than the length of the subject's original digital string, meaningless symbols are filled in front and back of the subject's answered digital string according to all possible filling conditions, the length is completed to be equal, and the edit distance is calculated with the subject's original digital string, and the minimum edit distance is taken as the final result.

[0011] As a preferred scheme of the memory ability evaluation method based on the edit distance, the fitting curve function is selected as the kernel function.

[0012] As a preferred scheme of the memory ability evaluation method based on the edit distance, due to the characteristics of memory, two special cases of complete memory and complete forgetting are included in the boundary conditions, that is:

[0013] f(0,0,0)=0

[0014] f(0,0,N)=1

[0015] Wherein, in the case of complete forgetting, all editing operations should be replaced; N is the bit length of the memory digital string.

[0016] As a preferred scheme of the memory ability evaluation method based on the edit distance, the expected condition comprises that the kernel function fits the curve to be fitted within a preset error allowable range, and the fitting result has high interpretability.

[0017] As a preferred scheme of the memory ability evaluation method based on the edit distance, the expected condition comprises that the kernel function fits the curve to be fitted within a preset error allowable range, and the fitting result has high interpretability.

[0018] The application is suitable for mixed memory task evaluation, and provides scientific theoretical support, better expands the stimulus type and evaluation mode of the n-back paradigm, and provides a new direction for working memory related research. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor. Among them:

[0020] Figure 1 The training process loss function change schematic diagram of the second embodiment of the application;

[0021] Figure 2 The fitting effect schematic diagram of the traditional technical scheme and the Ebbinghaus curve of the second embodiment of the application;

[0022] Figure 3 The fitting effect schematic diagram of the method and the Ebbinghaus curve of the second embodiment of the application. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the application.

[0024] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without the specific details set forth in this description. In other instances, well-known methods have not been described in detail in order not to unnecessarily obscure aspects of the present application.

[0025] Second, the "one embodiment" or "an embodiment" described herein as containing various features, structures, or characteristics can be combined with one or more other embodiments to form an embodiment of the application. The repetitions of "in one embodiment" or "in an embodiment" throughout the description are not to be construed as repetitions of a single embodiment or a single implementation of the present application, but are to be construed as repetitions of independent embodiments and implementations of the present application.

[0026] The present application is described in detail below in conjunction with the drawings, which show the preferred embodiments of the present application. In the drawings, the thickness of layers or regions can be exaggerated for clarity. The drawings are provided for purposes of illustration only and merely depict typical or example embodiments of the application. However, the application is not limited to the embodiments described herein, but can be practiced with modification and alteration within the scope of the present application. Furthermore, the present application covers and includes all suitable variations of the preferred embodiments as well as all suitable combinations of the elements of the preferred embodiments.

[0027] In the description of the present application, it should be noted that the terms "upper and lower", "inner and outer", and the like indicate the positional or directional relationship shown in the drawings, and are used only for the purpose of facilitating the description of the present application and simplifying the description, and thus should not be construed to indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the present application. In addition, the terms "first", "second", or "third" are used only for the purpose of description, and should not be construed as indicating or implying relative importance.

[0028] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection", and "connection" should be understood broadly, for example, can be fixed connection, detachable connection or integral connection; can also be mechanical connection, electrical connection or direct connection, can also be indirectly connected through an intermediate medium, or can be the internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0029] Embodiment 1

[0030] The present application provides a memory capacity evaluation method based on edit distance, comprising:

[0031] S1: Collect the original digital string of the subject's memory, the digital string answered by the subject, and the interval time.

[0032] It should be noted that in order to collect more data, the experiment needs to be designed before data collection; because the final fitting needs to be done by Ebbinghaus curve, the experiment design needs to be as close as possible to the original memory experiment of Ebbinghaus; for example, the experiment can adopt a number memory scheme: the subject needs to remember a string of meaningless numbers, such as 541862, etc., and after a certain time, the subject needs to recall the number string as much as possible according to memory and record his memory part.

[0033] In order to better fit the curve later, this embodiment is based on the key points of the original memory experiment of Ebbinghaus, and the interval time is selected, such as 5 minutes, 30 minutes, 12 hours, one day, etc.; the collected data includes the original number string remembered by the subject, the number string answered by the subject and the interval time.

[0034] S2: Calculate the shortest edit distance of the original number string remembered by the subject and the number string answered by the subject by using the edit distance algorithm.

[0035] It should be noted that the edit distance refers to the minimum number of editing operations required to convert one string into another; the greater the distance between them, the more different they are; the permitted editing operations include replacing one character with another, inserting a character, and deleting a character; edit distance is often used in plagiarism detection, spelling detection, etc.; it includes three editing operations with the same weight in statistics, and only the data of the minimum editing is counted, for example, when calculating the edit distance between 12345 and 13425, according to the minimum editing, 13425 is regarded as 12345 first deleting 2: 12345->1345, then adding 2 after 4: 1345->13425, a total of two-step editing, instead of three-step editing by replacing 234 with 342 in turn.

[0036] The original number string remembered by the subject and the number string answered by the subject are input into the edit distance algorithm to calculate the edit distance, which is written according to the principle of edit distance, and can calculate the shortest edit distance and the corresponding number of three kinds of editing.

[0037] For example: calculate the edit distance between 12345 and 13425, and get the result as the shortest edit distance: 2, wherein the replacement edit: 0 times, the insertion edit: 1 time, and the deletion edit: 1 time.

[0038] In order to better fit the actual memory situation, when calculating the edit distance:

[0039] (1) If the length of the number string answered by the subject is greater than the length of the original number string remembered by the subject, the original number string remembered by the subject will be edited with the length of the number string answered by the subject in turn, and the minimum edit distance is taken as the final result.

[0040] For example, if the original number string to be memorized is 12345 and the number string given by the subject is 123456, the edit distances of 12345-12345 and 12345-23456 are calculated in sequence, and the minimum result, i.e., the result of the first group, is taken as the final result.

[0041] (2) If the length of the number string given by the subject is less than the length of the original number string to be memorized, meaningless symbols are filled in front of and behind the number string given by the subject according to all possible filling conditions, the number string is completed to be equal in length, and the edit distances of the completed number string and the original number string to be memorized are calculated, and the minimum edit distance is taken as the final result.

[0042] For example, if the original number string to be memorized is 12345 and the number string given by the subject is 1234, meaningless symbols are filled in to obtain 12345-*1234 and 12345-1234*, and the minimum result, i.e., the result of the second group, is taken as the final result.

[0043] S3: Select a kernel function, calculate the forgetting information ratio through the kernel function, and obtain the data points in combination with the forgetting information ratio and the interval time.

[0044] Since the method needs to fit the function to the Ebbinghaus forgetting curve, and the meaning of the coordinate points of the forgetting curve is the forgetting information ratio after the current time interval, the kernel function of the embodiment is a model for expressing the forgetting information ratio using the edit distance.

[0045] The embodiment selects a fitting curve function as the kernel function. The fitting curve function can be divided into linear functions and logarithmic functions, exponential functions, etc. according to the function type. The embodiment selects the simplest linear function type as an example for illustration. Let the forgetting information ratio be f(x, y, z), then:

[0046]

[0047] In the formula, x, y, and z respectively represent the number of insertion, deletion, and replacement edits, a, b, and c are respectively the corresponding weight values, and N is the total information amount of the memory data represented by the original number string to be memorized. In the formula, N uses the length of the original number string to be memorized (herein, only a demonstrative kernel function is used).

[0048] If the formula is explained from the perspective of psychology, the formula divides the distortion of human memory information into three cases: information loss (deletion), information increase (insertion), and information error (replacement). Each case has a weight value representing the loss rate of the corresponding information when the case occurs, and the total information loss rate can be linearly represented by the three cases.

[0049] Further, the collected data is substituted into the selected kernel function to obtain a forgetting information ratio, and the forgetting information ratio and the interval time are combined to obtain data points, wherein the abscissa of the data points is the forgetting information ratio, and the ordinate is the interval time.

[0050] S4: The data points and the points on the Ebbinghaus forgetting curve are fitted by using a curve fitting algorithm.

[0051] The curve fitting algorithm can select some commonly used toolboxes, such as the cftool toolbox of Matlab, or select ant colony algorithm, genetic algorithm, etc. Some deep learning algorithms can also have good effects, such as linear neural network to fit linear function, etc.

[0052] After selecting the curve fitting algorithm, the data points and the points on the Ebbinghaus forgetting curve are fitted, and the fitting purpose is to find a certain kernel function to calculate the forgetting information ratio of a certain time interval.

[0053] According to the fitting result, the parameters of the kernel function are continuously adjusted. If the kernel function selected is the linear function mentioned in S3, the parameters a, b, and c are continuously adjusted according to the fitting result. The change of the parameters can change the form of the function, and the best parameter combination can make the form of the kernel function closest to the curve to be fitted, that is, make the fitted curve closer to the original Ebbinghaus curve.

[0054] In addition, due to the characteristics of memory, two special cases of complete memory and complete forgetting are included in the boundary conditions, that is:

[0055] f(0,0,0) = 0

[0056] f(0,0,N) = 1

[0057] Among them, in the case of complete forgetting, all editing operations should be replaced; N is the bit length of the memory number string.

[0058] S5: If the fitting result meets the expected condition, the parameters of the kernel function are adjusted according to the fitting result to obtain the best kernel function f(x, y, z); otherwise, the kernel function is reselected according to the reason for the fitting failure.

[0059] Expected conditions:

[0060] (1) The kernel function fits the curve to be fitted within the preset error allowed range;

[0061] (2) The fitting result has high interpretability.

[0062] For example, because the kernel function is a linear function, the weight of the three parameters should not be greater than 1, which means that the boundary condition f(0, 0, N) = 1 of the "completely forget" case will no longer hold, and the weights of the insertion and deletion should be equal to the replacement because from the perspective of editing operations, a replacement operation is actually a combination of insertion and deletion.

[0063] S6: Calculate the memory rate of the subject through the optimal kernel function to evaluate the memory ability.

[0064] Preferably, the optimal kernel function has great significance in psychological research and memory-related paradigm research. For example, the mechanism of working memory can be deeply researched through the function type and the size of each weight of the optimal kernel function, or a new working memory model can be proposed. Based on the n-back experimental paradigm, the mixed memory task form can also be designed through the function type and the size of each weight of the optimal kernel function.

[0065] Embodiment 2

[0066] In order to verify the technical effects adopted in the method, the traditional technical solution and the method are compared and tested in this embodiment to compare the test results by scientific means to verify the real effect of the method.

[0067] Traditional technical solution: In the memory span measurement experiment, the longest continuous sequence length that the subject can recall is used as the memory span to evaluate the memory ability of the subject. This method has a large error and is difficult to fit the Ebbinghaus forgetting curve.

[0068] In order to verify that the method has higher scientificity compared with the traditional technical solution, the memory data measured by the traditional technical solution and the method are used to calculate the memory rate in this embodiment, and the Ebbinghaus curve is compared.

[0069] Test environment: First, the memory experiment is designed, and the memory experiment of Ebbinghaus is referred to for design. In each trial, the subject is required to remember a meaningless word composed of 10 random letters. All words have no pronunciation rules to avoid logical association of the subject. The subject needs to remember the whole word by reading each letter at a certain frequency until the word can be completely recited. After that, the timing starts. During the timing, the subject is not allowed to repeat the word in the brain or any other way to review and strengthen the memory of the word. After the timing ends, the subject is required to reproduce the word and write the answer on the paper. The words the subject is required to remember, the words the subject reproduces, and the memory interval are collected as experimental data. A total of 40 groups of data are collected, and the memory interval is from 5s to 20min. The test samples collected by the above experiment are used to respectively use the traditional technical solution and the memory ability evaluation method based on the edit distance of the present application.

[0070] By using the method, the BP linear neural network written by python is used to fit the linear kernel function and the Ebbinghaus forgetting curve. The learning rate is 0.001, each data is repeated 1000 times, the average point line distance is calculated as the loss function, and the result is as shown in Figure 1 In order to improve the persuasiveness, 20 groups of data in the data with time interval of 5min to 20min are taken as the training set, and 20 groups of data with time interval of 5 to 30s are taken as the test set. After the corresponding kernel function is obtained by fitting, the test data is respectively brought into the kernel function and the traditional memory breadth calculation method to calculate the memory rate, and the Ebbinghaus forgetting curve corresponding to the time interval is drawn for comparison. According to the error between the memory rate calculated by the two methods and the corresponding value of the Ebbinghaus forgetting curve, the advantages and disadvantages of the method can be directly reflected, and the result is as shown in Figure 2 、 Figure 3 Figure 2 The fitting effect of the traditional technical solution and the Ebbinghaus curve, Figure 3 The fitting effect of the method and the Ebbinghaus curve, it can be seen that the fitting effect obtained by the method is obviously better than that of the traditional technical solution.

[0071] ​It should be appreciated that embodiments of the present application can be realized by computer hardware, a combination of hardware and software, or by computer instructions stored on a non-transitory computer-readable storage medium. The methods can be implemented in a computer program using standard programming techniques— including non-transitory computer-readable storage media configured with a computer program to implement the methods in which the storage media so configured causes a computer to operate in a specific and predefined manner— according to the methods described in the detailed embodiments and drawings. Each program can be implemented in a high level procedural or object-oriented programming language to communicate with a computer system. However, the programs can be implemented in assembly or machine language, if desired. In any case, the language can be a compiled or interpreted language. Also, the programs can be able to run on standalone systems or in conjunction with other programs or in a distributed computing environment, or any combination thereof. Further, the programs can be stored in any pure or mixed mode execution environment.

[0072] Further, the operations of the processes described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes described herein (or variations and / or combinations thereof) can be performed under the control of one or more computer systems configured with executable instructions (e.g., computer programs, one or more computer programs, or one or more applications) to perform the processes, by hardware, or combinations thereof. The computer programs include processor-executable instructions that are stored on a non-transitory computer-readable storage medium.

[0073] Further, the methods can be implemented in any suitable type of computing platform operably connected to, including but not limited to, a personal computer, mini-computer, mainframe, workstation, networked or distributed computing environment, separate or integrated computer platforms, or in communication with charged particle tools or other imaging devices, and the like. Aspects of the present application can be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated to the computing platform, such as a hard disk, optical read and / or write storage media, RAM, ROM, and the like, such that it can be read by a programmable computer to configure and operate the computer to perform the processes described herein. Further, the machine-readable code, or portions thereof, can be transmitted over wired or wireless networks. The present application described herein includes these and other different types of non-transitory computer-readable storage media when such media include instructions or programs to implement the steps described above in conjunction with a microprocessor or other data processor. The present application also includes the computer itself when programmed according to the methods and techniques described herein. The computer programs are able to apply to input data to perform the functions described herein, thereby transforming the input data to generate output data that is stored to non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In the preferred embodiments of the present application, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects produced on a display.

[0074] As used in this application, the terms "component," "module," "system" and the like are intended to refer to a computer-related entity, either hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and / or a computer. By way of illustration, both an application running on a computing device and the computing device can be a component. One or more components can reside within a process and / or thread of execution and a component can be localized, partially and / or entirely, in one computer or distributed between two or more computers. Also, these components can execute from various computer-readable media having various data structures stored thereon. The components can communicate by way of local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network such as the Internet with other systems via the signal).

[0075] It should be noted that the above-mentioned embodiments are only used to illustrate but not to limit the technical solutions of the present application, and although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalent replaced without departing from the spirit and scope of the technical solutions of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A method for evaluating a memory ability based on an edit distance, characterized by, The method comprises the following steps: Collecting the original number string memorized by the subject, the number string answered by the subject and the interval time; Calculating the shortest edit distance of the original number string memorized by the subject and the number string answered by the subject by using the edit distance algorithm; Selecting a kernel function, calculating the forgetting information ratio by the kernel function, combining the forgetting information ratio and the interval time to obtain data points; Fitting the data points and the points on the Ebbinghaus forgetting curve by using a curve fitting algorithm; If the fitting result meets the expected condition, adjusting the parameters of the kernel function according to the fitting result to obtain the best kernel function; otherwise, re-fitting / re-selecting the kernel function according to the reason for the fitting failure; Calculating the memory rate of the subject by the best kernel function to evaluate the memory ability; Wherein, the abscissa of the data point is the forgetting information ratio, and the ordinate is the interval time; The kernel function comprises: selecting a fitting curve function as the kernel function; Due to the characteristics of memory, two special cases of complete memory and complete forgetting are included in the boundary conditions, that is: f(0,0,0)＝0 f(0,0,N)=1 Wherein, in the case of complete forgetting, all editing operations should be replacement; N is the bit length of the memory number string.

2. The evaluation method of memory ability based on edit distance according to claim 1, wherein, The method comprises the following steps: Based on the key points of the original memory experiment of Ebbinghaus, the interval time is selected.

3. The evaluation method of memory ability based on edit distance according to claim 1 or 2, characterized in that, The method comprises the following steps: When calculating the edit distance, if the length of the number string answered by the subject is greater than the length of the original number string memorized by the subject, the original number string memorized by the subject will be sequentially calculated with the length equal to the length of the number string answered by the subject in the edit distance, and the minimum edit distance is taken as the final result.

4. The evaluation method of memory ability based on edit distance according to claim 2, wherein, The method comprises the following steps: When calculating the edit distance, if the length of the number string answered by the subject is less than the length of the original number string memorized by the subject, meaningless symbols will be filled in front and back of the number string answered by the subject according to all possible filling conditions to complete it to the same length, and the edit distance will be calculated with the original number string memorized by the subject, and the minimum edit distance is taken as the final result.

5. The evaluation method of memory ability based on edit distance according to claim 4, wherein, The expected conditions comprise: The kernel function fits the curve to be fitted within a predetermined error tolerance range; The fitting result has high interpretability.

6. The evaluation method of memory ability based on edit distance according to claim 5, wherein, The method comprises the following steps: Based on the n-back experiment paradigm, and through the function type and weight size of the best kernel function, a mixed memory task form is designed.

Images