A multimodal nursing service quality evaluation method and system based on data analysis

Abstract

The present invention discloses a multimodal nursing service quality evaluation method and system based on data analysis, which belongs to the field of nursing service technology. The abnormal state in the nursing service link is monitored in real time through a combination of multimodal sensors, the sensor data is timely processed, and a nursing service abnormal feedback feature label is constructed; based on the uniformly configured nursing port, service object and time node data, a port local data matrix is ​​established, and a feature portrait matrix is ​​generated by adjacent column difference and row difference calculation, and the abnormal score is realized by combining the modulus mean and modulus variance, and finally the nursing task arrangement is dynamically adjusted; the system includes a multimodal device module, a data preprocessing module, a data fitting module and a service evaluation module, and solves the problems of data dispersion and feedback lag in traditional nursing evaluation through sensor collaboration, matrix modeling and dynamic threshold judgment, and significantly improves the accuracy of abnormal detection and the intelligent level of nursing service quality management.

Description

Technical Field

[0001] The present invention relates to the technical field of nursing services, and in particular to a multimodal nursing service quality assessment method and system based on data analysis. Background Art

[0002] In the modern medical service system, the quality of nursing services is directly related to the patient's rehabilitation effect and medical satisfaction. With the continuous advancement of medical technology and the increasing diversification of patient needs, nursing services are no longer limited to traditional physical care, but have expanded to multiple dimensions such as psychology, nutrition, and rehabilitation, forming a new model of multimodal nursing services. However, the implementation of this multimodal nursing service model has also posed higher challenges to the evaluation of nursing quality.

[0003] Currently, the quality assessment of nursing services mainly relies on manual observation records or single sensor monitoring (such as heart rate monitoring bracelets). Manual assessment is highly subjective and has poor real-time performance, while a single sensor cannot fully capture multi-dimensional anomalies in nursing services (such as operational standardization, environmental risks, and coordinated changes in the patient's physiological state). Especially under the multimodal nursing service model, the nursing service links are more complicated and the equipment and technologies involved are more diverse. For example, whether the nurse's turning over operation is standardized requires the combination of visual motion capture and mattress pressure sensor data. Existing technologies make it difficult to achieve collaborative analysis of multimodal data and meet the needs of comprehensive, objective, and real-time assessment of nursing service quality. Summary of the invention

[0004] The purpose of the present invention is to provide a multimodal nursing service quality assessment method and system based on data analysis to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A multimodal nursing service quality evaluation system based on data analysis, the system comprises: a multimodal equipment module, a data preprocessing module, a data fitting module and a service evaluation module;

[0007] The multimodal device module is used to monitor and feedback abnormal conditions of the nursing service link and to time the monitoring feedback behavior of the multimodal sensor;

[0008] The data preprocessing module is used to uniformly configure the data, including the nursing port, service object, nursing service link, multimodal sensor, synchronization time node and timing trigger behavior, and is used to generate nursing service abnormal feedback feature labels;

[0009] The data fitting module establishes a port local data matrix based on the uniformly configured data, analyzes and fits the characteristic portrait of the caregiver when caring for the service object, and generates a characteristic portrait matrix;

[0010] The service evaluation module, based on the feature portrait matrix, performs anomaly scoring on the feature portraits generated by the caregivers when providing care services to the service recipients, and arranges the caregiver services according to the anomaly score values.

[0011] Further, the multimodal device module includes an abnormal feedback unit and a timing unit;

[0012] The abnormal feedback unit is used to configure a multimodal sensor, wherein a plurality of multimodal sensors are correspondingly set for one nursing service link, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label;

[0013] The timing unit is used to time the monitoring feedback behavior of the multimodal sensor and set a timed initialization feedback instruction, and the timed initialization feedback instruction is triggered when entering the starting stage of the nursing service. After the triggering, the time synchronization state of the multimodal sensor is synchronously updated to make the abnormal state of the monitoring feedback of the multimodal sensor have time synchronization.

[0014] Further, the data preprocessing module includes a data configuration unit and a data recording unit;

[0015] The data configuration unit is used to uniformly configure the data and form a modality set;

[0016] The data recording unit captures the abnormal state generated by the caregiver when providing nursing service to the object through the caregiver port, counts the abnormal state generated by each multimodal sensor in the modal set under the trigger tag, and records it in the nursing service abnormal feedback feature tag.

[0017] Further, the data fitting module includes a port local data matrix generation unit and a feature portrait generation unit;

[0018] The port local data matrix generating unit is used to establish a port local data matrix, wherein the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link;

[0019] The feature portrait generation unit, based on the port local data matrix, selects a matrix element as the matrix center element, calculates the adjacent column differences and row differences of the matrix element to form a center vector, and calculates the modulus of the center vector to generate a feature portrait matrix.

[0020] Further, the service evaluation module includes an abnormality evaluation unit and an early warning prompt unit;

[0021] The abnormality assessment unit, based on the feature portrait matrix, performs an abnormality score on the feature portrait generated by the care worker when caring for the service object to obtain an abnormality score value;

[0022] The early warning prompt unit is used to preset an abnormal threshold. If the abnormal score value is greater than or equal to the abnormal threshold, the nursing service task arrangement for the caregiver is stopped, and the caregiver is instructed to retrain the nursing service.

[0023] A multimodal nursing service quality assessment method based on data analysis, the method comprises the following steps:

[0024] Step S1: monitoring and feedback the abnormal state of the nursing service link through multimodal sensors, and the combined feedback of the multimodal sensors constitutes a modal set, and generating a nursing service abnormal feedback feature label; timing the monitoring feedback behavior of the multimodal sensors, so that the abnormal state of the multimodal sensor monitoring feedback has time synchronization;

[0025] Step S2: uniformly configure the data, including the care worker port, service object, care worker service link, multimodal sensor, synchronization time node and timely triggering behavior; capture the abnormal state generated by the care worker when caring for the service object through the care worker port, and count the abnormal state generated by each multimodal sensor in the modal set under the trigger tag to record it in the nursing service abnormal feedback feature tag;

[0026] Step S3: Based on the uniformly configured data, a port local data matrix is ​​established, and the characteristic portrait of the caregiver in nursing service objects is analyzed and fitted to generate a characteristic portrait matrix;

[0027] Step S4: Based on the feature portrait matrix, perform an abnormality score on the feature portrait generated by the caregiver when providing nursing services to the service recipient, and arrange the nursing service according to the abnormal score value.

[0028] Furthermore, the specific implementation process of step S1 includes:

[0029] Based on the nursing service link, a multimodal sensor is configured, wherein one nursing service link corresponds to a plurality of multimodal sensors, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label, wherein the multimodal sensor is used to monitor the nursing service behavior, and the abnormal state refers to the functional state fed back by the multimodal sensor when the nursing service behavior is at risk;

[0030] Based on the execution logic of the nursing service link, the monitoring feedback behavior of the multimodal sensor is timed, and a timed initialization feedback instruction is set. When entering the starting link of the nursing service, the timed initialization feedback instruction is triggered. After the triggering, the time synchronization state of the multimodal sensor is synchronously updated to make the abnormal state of the monitoring feedback of the multimodal sensor have time synchronization.

[0031] Furthermore, the specific implementation process of step S2 includes:

[0032] Configure the data uniformly and record any h-th care worker port as , and one caregiver is configured with one caregiver port, and any e-th service object is recorded as , record any y-th nursing service link as , any i-th multimodal sensor is recorded as , then the nursing service link is counted The multi-modal sensors configured in the , I represents the total number of multimodal sensors, and the time synchronization sequence of the timing is recorded as , represents the jth synchronization time node of the timing, J represents the total number of synchronization time nodes of the timing, and based on the triggered behavior of the timing initialization feedback instruction, a trigger tag is added to the timing synchronization sequence of the timing, which is recorded as ,and , f represents the number of triggers;

[0033] Through the nursing port Capturing the caregivers’ The abnormal state generated when the tag is triggered The following statistical mode set The abnormal state generated by each multimodal sensor in the nursing service is recorded in the abnormal feedback feature label In, and ,in, Indicates that the tag is triggered Nursing service link duration, and , and Respectively indicate that in the trigger tag Nursing service link The start synchronization time node and the end synchronization time node, , , Indicates that the tag is triggered The following statistical mode set The total number of abnormal states generated by each multimodal sensor in;

[0034] In the above method, a time synchronization instruction is triggered at the beginning of the nursing phase to dynamically align the sensor data stream and solve the problem of cross-modal timing deviation.

[0035] Furthermore, the specific implementation process of step S3 includes:

[0036] Establish a port local data matrix, the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link, then the nursing port Capturing the caregivers’ The port local data matrix generated when , and the port local data matrix The matrix element corresponding to the fth row and the yth column in is the feature label of the abnormal feedback of nursing service ;

[0037] Based on the port local data matrix, select the matrix elements is the center element of the matrix, and calculates the matrix element Adjacent column and row differences of :

[0038] ;

[0039] In the formula, Represents a matrix element The adjacent column differences of Represents a matrix element The adjacent row differences of ;

[0040] Based on matrix elements The adjacent column differences and row differences of form the central vector , calculate the modulus of the center vector , and the model Mapping to port local data matrix In the matrix position of row f and column y, we get the nursing staff’s nursing service object The feature image matrix at time , recorded as ;

[0041] In the above method, the caregiver behavior, service object status and sensor feedback are mapped into a matrix structure, and hidden abnormal patterns can be extracted through differential calculation.

[0042] Furthermore, the specific implementation process of step S4 includes:

[0043] Based on the feature profile matrix , for nursing staff in nursing service recipients The feature portrait generated at the time is scored for anomaly, and the anomaly score value is obtained , represents the modulus mean, represents the modular variance, and , Y represents the total number of nursing service links, and F represents the total number of triggered tags;

[0044] Preset an abnormal threshold. If the abnormal score is greater than or equal to the abnormal threshold, stop arranging nursing service tasks for the nurse and instruct the nurse to retrain the nursing service.

[0045] In the above method, the adaptive threshold is set in combination with the statistical distribution (mean value, variance) of the feature profile matrix to achieve multi-scale evaluation from single anomaly to long-term behavioral degradation.

[0046] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: in a multimodal nursing service quality evaluation method and system based on data analysis provided by the present invention, abnormal states in the nursing service link are monitored in real time through a combination of multimodal sensors, sensor data is processed in a timely manner, and a nursing service abnormal feedback feature label is constructed; based on uniformly configured nursing ports, service objects and time node data, a port local data matrix is ​​established, a feature portrait matrix is ​​generated by adjacent column differences and row differences, an abnormal score is achieved by combining the modulus mean and the modulus variance, and finally the nursing task arrangement is dynamically adjusted; the system includes a multimodal device module, a data preprocessing module, a data fitting module and a service evaluation module, and solves the problems of data dispersion and feedback lag in traditional nursing evaluation through sensor collaboration, matrix modeling and dynamic threshold judgment, and significantly improves the accuracy of abnormality detection and the intelligent level of nursing service quality management. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0048] Figure 1 It is a schematic diagram of the steps of a multimodal nursing service quality assessment method based on data analysis of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] In the first embodiment: a multimodal nursing service quality evaluation system based on data analysis is provided, the system comprising: a multimodal device module, a data preprocessing module, a data fitting module and a service evaluation module;

[0051] The multimodal device module is used to monitor and feedback abnormal conditions of the nursing service link and to time the monitoring feedback behavior of the multimodal sensor;

[0052] Wherein, the multimodal device module includes an abnormal feedback unit and a timing unit;

[0053] The abnormal feedback unit is used to configure a multimodal sensor, wherein a plurality of multimodal sensors are correspondingly set for one nursing service link, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label;

[0054] The timing unit is used to time the monitoring feedback behavior of the multimodal sensor and set a timed initialization feedback instruction, and the timed initialization feedback instruction is triggered when entering the starting link of the nursing service, and after the triggering, the time synchronization state of the multimodal sensor is synchronously updated, so that the abnormal state of the monitoring feedback of the multimodal sensor has time synchronization;

[0055] The data preprocessing module is used to uniformly configure the data, including the nursing port, service object, nursing service link, multimodal sensor, synchronization time node and timing trigger behavior, and is used to generate nursing service abnormal feedback feature labels;

[0056] Wherein, the data preprocessing module includes a data configuration unit and a data recording unit;

[0057] The data configuration unit is used to uniformly configure the data and form a modality set;

[0058] The data recording unit captures the abnormal state generated by the caregiver when the caregiver is providing care services to the object through the caregiver port, counts the abnormal state generated by each multimodal sensor in the modal set under the trigger tag, and records it in the nursing service abnormal feedback feature tag;

[0059] The data fitting module establishes a port local data matrix based on the uniformly configured data, analyzes and fits the characteristic portrait of the caregiver when caring for the service object, and generates a characteristic portrait matrix;

[0060] Wherein, the data fitting module includes a port local data matrix generation unit and a feature portrait generation unit;

[0061] The port local data matrix generating unit is used to establish a port local data matrix, wherein the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link;

[0062] The feature portrait generating unit, based on the port local data matrix, selects a matrix element as the matrix center element, calculates the adjacent column differences and row differences of the matrix element to form a center vector, and calculates the modulus of the center vector to generate a feature portrait matrix;

[0063] The service evaluation module, based on the feature portrait matrix, performs an abnormality score on the feature portrait generated by the caregiver when caring for the service object, and arranges the caregiver service according to the abnormal score value;

[0064] Wherein, the service evaluation module includes an abnormality evaluation unit and an early warning prompt unit;

[0065] The abnormality assessment unit, based on the feature portrait matrix, performs an abnormality score on the feature portrait generated by the care worker when caring for the service object to obtain an abnormality score value;

[0066] The early warning prompt unit is used to preset an abnormal threshold. If the abnormal score value is greater than or equal to the abnormal threshold, the nursing service task arrangement for the caregiver is stopped, and the caregiver is instructed to retrain the nursing service.

[0067] See also Figure 1 In the second embodiment, a multimodal nursing service quality assessment method based on data analysis is provided to be applicable to the first embodiment, and the method comprises the following steps:

[0068] Step S1: monitoring and feedback the abnormal state of the nursing service link through multimodal sensors, and the combined feedback of the multimodal sensors constitutes a modal set, and generating a nursing service abnormal feedback feature label; timing the monitoring feedback behavior of the multimodal sensors, so that the abnormal state of the multimodal sensor monitoring feedback has time synchronization;

[0069] Exemplarily, based on the nursing service link, a multimodal sensor is configured, wherein one nursing service link corresponds to a plurality of multimodal sensors, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label, wherein the multimodal sensor is used to monitor the nursing service behavior, and the abnormal state refers to the functional state fed back by the multimodal sensor when the nursing service behavior is at risk;

[0070] Based on the execution logic of the nursing service link, the monitoring feedback behavior of the multimodal sensor is timed, and a timed initialization feedback instruction is set. When entering the starting link of the nursing service, the timed initialization feedback instruction is triggered. After the triggering, the time synchronization state of the multimodal sensor is synchronously updated to make the abnormal state of the monitoring feedback of the multimodal sensor have time synchronization.

[0071] Step S2: uniformly configure the data, including the care worker port, service object, care worker service link, multimodal sensor, synchronization time node and timely triggering behavior; capture the abnormal state generated by the care worker when caring for the service object through the care worker port, and count the abnormal state generated by each multimodal sensor in the modal set under the trigger tag to record it in the nursing service abnormal feedback feature tag;

[0072] For example, the data is uniformly configured, and any h-th care worker port is recorded as , and one caregiver is configured with one caregiver port, and any e-th service object is recorded as , record any y-th nursing service link as , any i-th multimodal sensor is recorded as , then the nursing service link is counted The multi-modal sensors configured in the , I represents the total number of multimodal sensors, and the time synchronization sequence of the timing is recorded as , represents the jth synchronization time node of the timing, J represents the total number of synchronization time nodes of the timing, and based on the triggered behavior of the timing initialization feedback instruction, a trigger tag is added to the timing synchronization sequence of the timing, which is recorded as ,and , f represents the number of triggers;

[0073] Through the nursing port Capturing the caregivers’ The abnormal state generated when the tag is triggered The following statistical mode set The abnormal state generated by each multimodal sensor in the nursing service is recorded in the abnormal feedback feature label In, and ,in, Indicates that the tag is triggered Nursing service link duration, and , and Respectively indicate that in the trigger tag Nursing service link The start synchronization time node and the end synchronization time node, , , Indicates that the tag is triggered The following statistical mode set The total number of abnormal states generated by each multimodal sensor in;

[0074] In the fall detection link, the sensor modal combination involves visual sensors, cameras to capture changes in body posture (such as body tilt angle > 45 degrees), tactile sensors, wearable device accelerometers to detect severe vibrations, voice sensors, microphones to capture cries for help or collisions, and physiological sensors, heart rate sensors to show a sudden increase in heart rate (such as > 120 beats / minute);

[0075] In the operation specification link, it involves the combination of sensor modes, visual sensors, cameras to detect omissions in disinfection steps (such as direct contact with patients without washing hands), tactile sensors, smart gloves to monitor the strength of operations (such as excessive force in turning over), and voice sensors, microphones to analyze the tone of conversation between nurses and patients (such as the use of inappropriate language);

[0076] In the response delay link, the modal combination of sensors is involved, including visual sensors, mattress pressure sensors to record the turning interval (such as no turning over for more than 2 hours), environmental sensors, timers and motion sensors (such as failure to feed medicine on time), so as to monitor whether the caregivers complete the necessary operations within the specified time.

[0077] Step S3: Based on the uniformly configured data, a port local data matrix is ​​established, and the characteristic portrait of the caregiver in nursing service objects is analyzed and fitted to generate a characteristic portrait matrix;

[0078] Exemplarily, a port local data matrix is ​​established, the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link, then the nursing port Capturing the caregivers’ The port local data matrix generated when , and the port local data matrix The matrix element corresponding to the fth row and the yth column in is the feature label of the abnormal feedback of nursing service ;

[0079] Based on the port local data matrix, select the matrix elements is the center element of the matrix, and calculates the matrix element Adjacent column and row differences of :

[0080] ;

[0081] In the formula, Represents a matrix element The adjacent column differences of Represents a matrix element The adjacent row differences of ;

[0082] Based on matrix elements The adjacent column differences and row differences of form the central vector , calculate the modulus of the center vector , and the model Mapping to port local data matrix In the matrix position of row f and column y, we get the nursing staff’s nursing service object The feature image matrix at time , recorded as .

[0083] Step S4: Based on the feature portrait matrix, the feature portraits generated by the caregivers when caring for the service recipients are scored for abnormality, and the caregiver service is arranged according to the abnormal score values;

[0084] For example, based on the feature profile matrix , for nursing staff in nursing service recipients The feature portrait generated at the time is scored for anomaly, and the anomaly score value is obtained , represents the modulus mean, represents the modular variance, and , Y represents the total number of nursing service links, and F represents the total number of triggered tags;

[0085] An abnormal threshold is preset. If the abnormal score value is greater than or equal to the abnormal threshold, the nursing service task arrangement for the nurse is stopped, and the nurse is instructed to retrain the nursing service.

[0086] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0087] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multimodal nursing service quality assessment method based on data analysis, characterized in that: The method comprises the following steps: Step S1: monitoring and feedbacking abnormal states of nursing service links through multimodal sensors, timing the monitoring and feedback behaviors of the multimodal sensors, so that the abnormal states monitored and feedback by the multimodal sensors have time synchronization; Step S2: uniformly configure the data and record any h-th care worker port as W h , and one caregiver is configured with one caregiver port, and any e-th service object is recorded as S e , record any y-th nursing service link as X y , denote any i-th multimodal sensor as M i , then the nursing service link X y The multi-modal sensors configured in the y )={M i |i∈[1,I]}, I represents the total number of multimodal sensors, and the time synchronization sequence is recorded as A={t j |j∈[1,J]},t j represents the jth synchronization time node of the timing, J represents the total number of synchronization time nodes of the timing, and based on the triggered behavior of the timing initialization feedback instruction, a trigger tag is added to the timing synchronization sequence of the timing, which is recorded as A f , and A f =A, f represents the number of trigger times; Port W by carer h Capture the nursing staff's nursing service object e The abnormal state generated when triggering tag A f The statistical modal set O(X y ) and record the abnormal state generated by each multimodal sensor in the nursing service abnormal feedback feature label A f :X y In, and A f :X y =[T(A f |X y ), K(A f |X y )], where T(A f |X y ) indicates that when triggering tag A f Nursing service linkX y The duration of T(A f |X y )=t″-t′, t″ and t′ represent the trigger tag A f Nursing service linkX y The starting synchronization time node and the ending synchronization time node, t″∈A, t′∈A, K(A f |X y ) indicates that when triggering tag A f The statistical modal set O(X y ) The total number of abnormal states generated by each multimodal sensor; Step S3: Establish a port local data matrix, the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link. h Capture the nursing staff's nursing service object e The port local data matrix generated when e |W h ), and the port local data matrix R(S e |W h ) in which the matrix element corresponding to the fth row and the yth column is the feature label A of the abnormal feedback of nursing service f :X y ; Based on the port local data matrix, select the matrix element A f :X y is the center element of the matrix, calculate the matrix element A f :X y Adjacent column and row differences of : In the formula, L1(A f :X y ) represents the matrix element A f :X y The adjacent column difference, L2(A f :X y ) represents the matrix element A f :X y The adjacent row differences of ; Based on the matrix element A f :X y The adjacent column differences and row differences of form the central vector V(A f :X y )=(L1(A f :X y ), L2(A f :X y )), calculate the modulus of the center vector ‖V(A f :X y )‖, and the model ‖V(A f :X y )‖ is mapped to the port local data matrix R(S e |W h ) in the matrix position of the fth row and the yth column, we get the position of the care worker in the nursing service object S e The feature image matrix at this time is recorded as r(S e |W h ); Step S4: Based on the feature portrait matrix, perform an abnormality score on the feature portrait generated by the caregiver when providing nursing services to the service recipient, and arrange the nursing service according to the abnormal score value.

2. A multimodal nursing service quality assessment method based on data analysis according to claim 1, characterized in that: The specific implementation process of step S1 includes: Based on the nursing service link, a multimodal sensor is configured, wherein one nursing service link corresponds to a plurality of multimodal sensors, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label, wherein the multimodal sensor is used to monitor the nursing service behavior, and the abnormal state refers to the functional state fed back by the multimodal sensor when the nursing service behavior is at risk; Based on the execution logic of the nursing service link, the monitoring feedback behavior of the multimodal sensor is timed, and a timed initialization feedback instruction is set. When entering the starting link of the nursing service, the timed initialization feedback instruction is triggered. After the triggering, the time synchronization state of the multimodal sensor is synchronously updated to make the abnormal state of the monitoring feedback of the multimodal sensor have time synchronization.

3. A multimodal nursing service quality assessment method based on data analysis according to claim 1, characterized in that: The specific implementation process of step S4 includes: Based on the feature portrait matrix r(S e |W h ), for nursing staff in nursing service recipients S e The feature portrait generated at the time is scored for anomaly, and the anomaly score value is obtained μ represents the modulus mean, σ 2 represents the modular variance, and Y represents the total number of nursing service sessions, and F represents the total number of triggered tags; An abnormal threshold is preset. If the abnormal score value is greater than or equal to the abnormal threshold, the nursing service task arrangement for the nurse is stopped, and the nurse is instructed to retrain the nursing service.

4. A multimodal nursing service quality evaluation system based on data analysis, executing a multimodal nursing service quality evaluation method based on data analysis as described in any one of claims 1 to 3, characterized in that: The system includes: a multimodal device module, a data preprocessing module, a data fitting module and a service evaluation module; The multimodal device module is used to monitor the abnormal state of the feedback nursing service link, and time the monitoring feedback behavior of the multimodal sensor, and based on the triggered behavior of the timed initialization feedback instruction, add a trigger tag to the time synchronization sequence of the timed time; The data preprocessing module is used to uniformly configure the data, including the nursing port, service object, nursing service link, multimodal sensor, synchronization time node and timing trigger behavior, and is used to generate the abnormal feedback feature label of the nursing service; the data fitting module is used to establish the port local data matrix based on the uniformly configured data, analyze and fit the feature portrait of the nursing worker when providing nursing service to the object, and generate the feature portrait matrix; The service evaluation module, based on the feature portrait matrix, performs anomaly scoring on the feature portraits generated by the caregivers when providing care services to the service recipients, and arranges the caregiver services according to the anomaly score values.

5. A multimodal nursing service quality evaluation system based on data analysis according to claim 4, characterized in that: The multimodal device module includes an abnormality feedback unit and a timing unit; The abnormal feedback unit is used to configure a multimodal sensor, wherein a plurality of multimodal sensors are correspondingly set for one nursing service link, and the abnormal state monitored and fed back by the multimodal sensor is separated to generate a nursing service abnormal feedback feature label; The timing unit is used to time the monitoring feedback behavior of the multimodal sensor and set a timed initialization feedback instruction, and the timed initialization feedback instruction is triggered when entering the starting stage of the nursing service. After the triggering, the time synchronization state of the multimodal sensor is synchronously updated to make the abnormal state of the monitoring feedback of the multimodal sensor have time synchronization.

6. A multimodal nursing service quality evaluation system based on data analysis according to claim 4, characterized in that: The data preprocessing module includes a data configuration unit and a data recording unit; The data configuration unit is used to uniformly configure the data and form a modality set; The data recording unit captures the abnormal state generated by the caregiver when providing nursing service to the object through the caregiver port, counts the abnormal state generated by each multimodal sensor in the modal set under the trigger tag, and records it in the nursing service abnormal feedback feature tag.

7. A multimodal nursing service quality evaluation system based on data analysis according to claim 4, characterized in that: The data fitting module includes a port local data matrix generation unit and a feature image generation unit; The port local data matrix generating unit is used to establish a port local data matrix, wherein the row number of the port local data matrix is ​​the number of times the tag is triggered, and the column number of the port local data matrix is ​​the index number of the nursing service link; The feature portrait generation unit, based on the port local data matrix, selects a matrix element as the matrix center element, calculates the adjacent column differences and row differences of the matrix element to form a center vector, and calculates the modulus of the center vector to generate a feature portrait matrix.

8. A multimodal nursing service quality evaluation system based on data analysis according to claim 4, characterized in that: The service evaluation module includes an abnormality evaluation unit and an early warning prompt unit; The abnormality assessment unit, based on the feature portrait matrix, performs an abnormality score on the feature portrait generated by the care worker when caring for the service object to obtain an abnormality score value; The early warning prompt unit is used to preset an abnormal threshold. If the abnormal score value is greater than or equal to the abnormal threshold, the nursing service task arrangement for the caregiver is stopped, and the caregiver is instructed to retrain the nursing service.

Images