CN116312936A - A method and device for designing software for assisting the use of cognitive digital medicines - Google Patents

Abstract

The invention relates to the technical field of information, in particular to a design method and a device of software for assisting in using a cognitive digital drug, wherein the design method of the software for assisting in using the cognitive digital drug comprises the following steps: importing doctor prescriptions and personal information of patients to obtain prescription data and personal information data; processing data information generated by the digital medicine to obtain initial data, and extracting features according to the initial data to obtain first feature data; sending a medication reminder to the patient according to the initial data and the prescription data; deep learning is carried out through a convolutional neural network according to the first characteristic data, and second characteristic data are obtained; pushing auxiliary treatment information to the patient according to the second characteristic data and the personal information data; based on the second characteristic data and the prescription information data, a patient motivation task is formulated, and the medication effect and experience of the digital medicine are improved.

Description

A method and device for designing software for assisting the use of cognitive digital medicines

technical field

The invention relates to the field of information technology, in particular to a software design method and device for assisting the use of cognitive digital medicine.

Background technique

Digital medicine refers to digital products that have similar treatment indications, treatment goals and expected treatment effects as traditional medicines (or can improve treatment efficiency and effectiveness through innovative technologies). Compared with traditional drugs, digital drugs have great advantages such as easy and fun, high compliance, timely statistical feedback of diagnosis and treatment data, real-time automatic optimization of diagnosis and treatment plans based on artificial intelligence, and no pharmacological side effects. Referring to the research and development process of traditional drugs in the field of cognitive impairment, based on the existing technical foundation, the research and development model of digital drugs for cognitive impairment can be designed to provide reference guidance for the research and development of digital drugs in this field in the future, thereby accelerating Research and development of digital medicines in this field.

How to properly embed the auxiliary software of "digital medicine" into "digital medicine" and stimulate the auxiliary software in "digital medicine" to work normally through reasonable biological stimulation has become the research direction of current attention. In the existing digital drug assistant software, there are deficiencies in helping patients and doctors to grasp the patient's physical condition in real time. In the process of using drugs, they cannot effectively help and reward patients in real time, resulting in poor drug effects and poor treatment experience. Difference.

Contents of the invention

The embodiment of the present invention provides a software design method and device for assisting the use of cognitive digital medicines. Described technical scheme is as follows:

In one aspect, there is provided a method for designing cognitive digital drug assistance software, the method is implemented by an electronic device, and the method includes:

Import doctor's prescription and patient's personal information, obtain prescription data and personal information data.

The data information generated by the digital medicine is processed to obtain initial data, and feature extraction is performed according to the initial data to obtain first feature data.

According to the initial data and the prescription data, a medication reminder is sent to the patient.

The second feature data is obtained by performing deep learning through a convolutional neural network according to the first feature data.

According to the second characteristic data and the personal information data, auxiliary treatment information is pushed to the patient.

A patient motivational task is formulated based on the second characteristic data and the prescribing information data.

Optionally, the importing of doctor's prescription and patient's personal information to obtain prescription data and personal information data includes:

By using the wireless communication method of B/S architecture mode for data transmission, the personal information data and prescription data are sent from the hospital prescription server to the patient's mobile phone terminal, and the personal information data and prescription data are stored in SQLite.

Optionally, the initial data is obtained by processing the data information generated by the use of digital medicines, and feature extraction is performed according to the initial data to obtain the first feature data, including:

The data information of the patient's medication situation is transmitted through wireless communication, and the initial data is obtained through the data conversion interface between the digital drug and the platform to perform filtering and information model conversion operations, and the feature extraction is performed based on the initial data using the sorting condition mutual information method , to obtain the first feature data.

Optionally, performing deep learning through a convolutional neural network according to the first feature data to obtain second feature data includes:

Based on the first feature data, a dialogue generation model is established through a neural network-based word vector representation; by introducing the Seq2seq neural network model of the attention mechanism, the input first feature data is converted into a computer binary code; according to the computer Binary encoding is based on convolutional neural network for deep learning to obtain second feature data.

Optionally, said formulating a patient incentive task based on said second characteristic data includes:

The memory state of the patient under the current test environment is obtained, and based on the second feature data and the prescription data, dynamic matching of incentive tasks is performed according to the mental state of the patient through the learning method of the long-short-term memory network.

In another aspect, a software design device for assisted use of cognitive digital medicine is provided, the device is applied to a software method for assisted use of cognitive digital medicine, and the device includes:

The information collection module is used to import doctor's prescription and patient's personal information, and obtain prescription data and personal information data.

The feature extraction module is used to process the data information generated by the digital medicine to obtain initial data, and perform feature extraction according to the initial data to obtain the first feature data.

The drug reminder module is configured to send medication reminders to patients according to the initial data and the prescription data.

A deep learning module, configured to perform deep learning through a convolutional neural network according to the first feature data to obtain second feature data.

The virtual assistant module is configured to push auxiliary treatment information to the patient according to the second characteristic data and the personal information data.

An incentive system module, configured to formulate patient incentive tasks based on the second feature data and the prescription information data.

Optionally, the medicine reminder module is further used for:

By using the wireless communication method of B/S architecture mode for data transmission, the personal information data and prescription data are sent from the hospital prescription server to the patient's mobile phone terminal, and the personal information data and prescription data are stored using SQL i te.

Optionally, the medicine reminder module is further used for:

The data information of the patient's medication situation is transmitted through wireless communication, and the initial data is obtained through the data conversion interface between the digital drug and the platform to perform filtering and information model conversion operations, and the feature extraction is performed based on the initial data using the sorting condition mutual information method , to obtain the first feature data.

Optionally, the virtual assistant module is further used for:

Based on the first feature data, a dialogue generation model is established through a neural network-based word vector representation to solve the mapping problem of the cyclic neural network between sequences; by introducing the Seq2seq neural network model of the attention mechanism, the input first feature The data is converted into a computer binary code; according to the computer binary code, deep learning is performed based on a convolutional neural network to obtain second feature data.

Optionally, the incentive system module is further used for:

The memory state of the patient under the current test environment is obtained, and based on the second feature data and the prescription data, dynamic matching of incentive tasks is performed according to the mental state of the patient through the learning method of the long-short-term memory network.

In another aspect, an electronic device is provided, the electronic device includes a processor and a memory, at least one instruction is stored in the memory, and the at least one instruction is loaded and executed by the processor to realize the above recognition Awareness of digital drug assisted use software design methods.

In another aspect, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, and the at least one instruction is loaded and executed by a processor to implement the above-mentioned software design method for assisting use of cognitive digital medicine .

The beneficial effects brought by the technical solutions provided by the embodiments of the present invention at least include:

The invention properly embeds the auxiliary software of "digital medicine" into the "digital medicine", and stimulates the auxiliary software in the "digital medicine" to work normally through reasonable interactive behavior, and collects and transmits the patient's medication data and human brain electrophysiology The data is transmitted to the server platform through a certain transmission method, the data is analyzed and processed, and guiding opinions and suggestive suggestions are made for the patient's medical rehabilitation. Help patients and doctors to grasp the patient's physical condition in real time, give real-time help and rewards to patients in the process of using drugs, and improve the effect and experience of medication.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

Fig. 1 is a flow chart of a software design method for assisting the use of cognitive digital medicine provided by an embodiment of the present invention;

Fig. 2 is a block diagram of a software design device for assisted use of cognitive digital medicine provided by an embodiment of the present invention;

Fig. 3 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.

Detailed ways

In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will describe in detail with reference to the drawings and specific embodiments.

An embodiment of the present invention provides a software design method for assisting the use of cognitive digital medicine, the method can be implemented by an electronic device, and the electronic device can be a terminal or a server. As shown in Figure 1, a flow chart of a method for designing cognitive digital drug assistance software, the processing flow of the method may include the following steps:

S1. Import doctor's prescription and patient's personal information, and obtain prescription data and personal information data.

Optionally, by using the B/S architecture mode for data transmission, the personal information data and prescription data are sent from the hospital prescription server to the patient's mobile phone, and the personal information data and prescription data are stored using SQL i te.

In a feasible implementation manner, the mobile terminal submits a request through URLConnect to obtain the data of the hospital prescription server. First call the openConnect ion method of a URL object to create a URLConnect object; then set the URLConnect parameters and POST request attributes; secondly obtain the output stream corresponding to the URLConnect instance to send the request parameters; finally read the data of the hospital prescription server.

S2. Process the data information generated by the digital medicine to obtain initial data, perform feature extraction according to the initial data, and obtain first feature data.

Optionally, the initial data is obtained by processing the data information generated by the use of digital medicines, and feature extraction is performed based on the initial data to obtain the first feature data, including:

The data information of the patient's medication status is transmitted through wireless communication, and the initial data is obtained through filtering and information model conversion operations through the data conversion interface between the digital drug and the platform. first feature data.

In a feasible implementation mode, the data information of the patient's medication situation is transmitted through Bluetooth or Wifi, and a data conversion interface is designed to connect the digital medicine with the platform, and the relevant data of the patient's medication situation is converted into a suitable After the data model, process and analyze again. Among them, the sorting conditional mutual information method is used for feature extraction on the patient's brain electrophysiological signal, and the sorting conditional mutual information method is used for feature extraction.

S3. According to the initial data and the prescription data, a medication reminder is sent to the patient.

In a feasible implementation, refer to the doctor to issue a digital drug prescription, check the details of the day's medication information, add and update reminders, and view historical records. Provide relevant video, picture and text information, community service communication and electronic calendar, medication compliance and drug statistics services, and long-term health education. Compatible with both IOS and Android platforms, users can log in through their mobile phone number, or through WeChat and QQ third-party software authorization to log in. After logging in, the account information can be changed, and new needs and feedback can be provided through the user feedback interface. problem communication. Build a "cloud + terminal" architecture, realize the multi-dimensional sharing of data and doctors, improve the efficiency of medication, and strengthen the effect of medication.

S4. Perform deep learning through a convolutional neural network according to the first feature data to obtain second feature data.

Optionally, performing deep learning through a convolutional neural network according to the first feature data to obtain second feature data, including:

Based on the first feature data, a dialogue generation model is established through neural network-based word vector representation; the input first feature data is converted into computer binary code by introducing the Seq2seq neural network model of attention mechanism; The product neural network is used for deep learning to obtain the second feature data.

In a feasible implementation, deep learning is to learn the internal laws and representation levels of sample data. The interpretation of information obtained in the learning process is of great help. Deep learning can learn details that people cannot observe in the data, and can Help us extract features from the data. The auxiliary software can use intelligent customized algorithms to serve the treatment process of rehabilitation patients based on patient information and medication conditions.

The training process of convolutional neural network includes two stages of forward propagation and back propagation. The forward propagation process mainly includes: convolution feature extraction, pooling and error calculation. The backpropagation process mainly includes: error feedback and weight update. The weights are initialized by random assignment; the information is then passed to the convolutional layer, the pooling layer, and the fully connected layer in turn, where the convolutional layer and the pooling layer can extract the most significant features of the observation data through a filter , and can also extract richer patient feature information by stacking multiple convolutional layers and pooling layers; transform and calculate the information of multiple hidden layers in the fully connected layer, and pass it to the output layer; finally, the actual The output result is compared with the expected output result, and if the error function meets the precision requirement, the result is output directly. If it is not satisfied, backpropagate the bias and weight back to update the weight until the weight tends to be stable to obtain the second characteristic data.

S5. According to the second feature data and personal information data, push auxiliary treatment information to the patient.

In a feasible implementation, the scale measurement tool is used in the dialogue to facilitate the development of targeted interventions, and on the other hand, it is necessary to understand user preferences in order to provide alternatives for them. In order to obtain the activity and emotional state of the elderly, technologies such as camera monitoring and emotion recognition can be used for identification and feedback. For patients at home, the most convenient service of voice assistants is to obtain answers to some medical questions. By obtaining sufficient authoritative health knowledge from professional channels, the knowledge base of voice assistants is carefully created to ensure that elderly patients with cognitive impairment can obtain exact answer. In order to deepen the information memory of elderly users, in addition to fixed-point push, the system can also start SMS notifications and send text content to the mobile phones of elderly patients. The records of health status are sent to users in the form of periodic reports, allowing users to clarify their own health status. At the same time, the assistant should also warn of problems found and provide corresponding adjustment plans. When patients have more pressing concerns, they need to be assured of easy access to family and other medical helplines. In addition, after the plan is formulated, it may be difficult to ensure compliance only by the voice supervision, reminder and encouragement of the voice assistant. Therefore, it is necessary to regularly send the patient's progress to the patient's family members, and encourage the family members to set up a health plan together to strengthen the relationship between patients and their families. emotional connection, and then improve the patient's self-efficacy, and ultimately achieve the purpose of optimizing behavioral habits.

S6. Based on the second feature data and the prescription information data, formulate a patient motivation task.

Optionally, based on the second characteristic data, a patient motivation task is formulated, including:

Obtain the memory state of the patient in the current test environment, based on the second feature data and prescription data, and use the learning method of the long-term short-term memory network to dynamically match the incentive task according to the mental state of the patient.

In a feasible implementation, the Long Short Term Memory network (Long Short Term Memory networks, LSTM) is composed of repeated units. Inside each unit, there are mainly forget gates, input gates and output gates to control. These units receive input from previous The input of a unit and the prescription data input x _t of the doctor at the current time step t. Each LSTM cell contains a cell state _ct and hidden state _ht , which are modulated by 4 neural network layers that control the flow of information into and out of the cell's memory.

LSTM first needs to calculate the current input x _t and the hidden layer output h _t-1 of the previous section through the forget gate. The formula for controlling the LSTM forget gate is as follows (23):

f _t =σ(W _f x _t +U _f h _t-1 +b _f )...(23)

Among them, W _f represents the weight matrix that maps the hidden layer input to the forget gate, and represents the weight matrix that connects the output state of the previous moment to the forget gate, b _f represents the bias vector, and σ represents the activation function. In the present invention, it is si gmod activation function. The forget gate t _f controls the degree of influence of the previous storage unit c _t-1 on the current storage unit c _t through the sigmod activation function, and discards redundant information. The forget gate will read and output a value between 0 and 1 through the gate. 1 means "completely reserved", 0 means "completely discarded".

Add new information to a single cell, the input gate controls the degree of influence of the input x _t and h _t-1 on the current storage unit, the sigmod of the input layer determines the information that needs to be updated, the tanh layer obtains an update content c _t , and c _t-1 to update.

Finally, the output gate can control the influence of the current unit c _t on the hidden state unit h _t , use the output gate o _t to filter the unit state, and use it for hidden state update to obtain the final output of the LSTM unit, according to the final output of the patient Incentive tasks for dynamic matching.

This part focuses on the task reward system of the auxiliary software, accepts the data processed by the virtual assistant twice, and adopts automatic task formulation for each stage of the patient's medication, and generates a task suitable for the patient. During the daily training tasks, calculate the game difficulty, game completion time and game ranking points rules to obtain relevant points. When patients complete the tasks or log in to the training system for daily check-in, the patient's incentive points can also be increased. At the same time, the interactive interface of the point mall is designed and completed. The incentive points obtained by patients can be used to redeem vouchers in the point mall or unlock some functions of the virtual assistant. After exiting the training system, the incentive system will automatically update the training content and check-in calendar.

The invention properly embeds the auxiliary software of "digital medicine" into the "digital medicine", and stimulates the auxiliary software in the "digital medicine" to work normally through reasonable interactive behavior, and collects and transmits the patient's medication data and human brain electrophysiology The data is transmitted to the server platform through a certain transmission method, the data is analyzed and processed, and guiding opinions and suggestive suggestions are made for the patient's medical rehabilitation. Help patients and doctors to grasp the patient's physical condition in real time, give real-time help and rewards to patients in the process of using drugs, and improve the effect and experience of medication.

Fig. 2 is a block diagram of a software design device for assisting use of cognitive digital medicine according to an exemplary embodiment, and the device is applied to implement a software design method for assisting use of cognitive digital medicine based on a time-varying graph. Referring to Figure 2, the device includes:

The information collection module is used to import doctor's prescription and patient's personal information, and obtain prescription data and personal information data.

The feature extraction module is used to process the data information generated by the digital medicine to obtain initial data, perform feature extraction according to the initial data, and obtain first feature data.

The drug reminder module is configured to send medication reminders to patients according to the first feature data and prescription data.

The deep learning module is used to perform deep learning through the convolutional neural network according to the first feature data to obtain the second feature data.

The virtual assistant module is used to push auxiliary treatment information to the patient according to the second feature data and personal information data.

An incentive system module, configured to formulate patient incentive tasks based on the second characteristic data and the prescription information data.

Optionally, the information collection module is further used for:

By using the wireless communication method of B/S architecture mode for data transmission, the personal information data and prescription data are sent from the hospital prescription server to the patient's mobile phone terminal, and the personal information data and prescription data are stored in SQLite.

Optionally, the feature extraction module is further used for:

The data information of the patient's medication status is transmitted through wireless communication, and the initial data is obtained through filtering and information model conversion operations through the data conversion interface between the digital drug and the platform. first feature data.

Optionally, the deep learning module is further used to:

Based on the first feature data, a dialogue generation model is established through a neural network-based word vector representation to solve the mapping problem of the cyclic neural network between sequences; the input first feature data is converted by introducing the Seq2seq neural network model of the attention mechanism It is a computer binary code; performing deep learning based on a convolutional neural network according to the computer binary code to obtain the second feature data.

Optionally, the incentive system module is further used to:

Obtain the memory state of the patient in the current test environment, based on the second feature data and prescription data, and use the learning method of the long-term short-term memory network to dynamically match the incentive task according to the mental state of the patient.

The invention properly embeds the auxiliary software of "digital medicine" into the "digital medicine", and stimulates the auxiliary software in the "digital medicine" to work normally through reasonable interactive behavior, and collects and transmits the patient's medication data and human brain electrophysiology The data is transmitted to the server platform through a certain transmission method, the data is analyzed and processed, and guiding opinions and suggestive suggestions are made for the patient's medical rehabilitation. Help patients and doctors to grasp the patient's physical condition in real time, give real-time help and rewards to patients in the process of using drugs, and improve the effect and experience of medication.

FIG. 3 is a schematic structural diagram of an electronic device 300 provided by an embodiment of the present invention. The electronic device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing ing un its , CPU) 301 and one or more memories 302, wherein at least one instruction is stored in the memory 302, and the at least one instruction is loaded and executed by the processor 301 to realize the above-mentioned cognitive digital medicine assistance Steps in using the software design method.

In an exemplary embodiment, there is also provided a computer-readable storage medium, such as a memory including instructions, which can be executed by a processor in the terminal to implement the above-mentioned method for designing cognitive digital medicine assistance software. For example, the computer readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like.

Those of ordinary skill in the art can understand that all or part of the steps for implementing the above embodiments can be completed by hardware, and can also be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium. The above-mentioned The storage medium mentioned may be a read-only memory, a magnetic disk or an optical disk, and the like.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (10)

1. A method for designing a cognitive digital drug adjunctive use software, the method comprising:

importing doctor prescriptions and personal information of patients to obtain prescription data and personal information data;

processing data information generated by the digital medicine to obtain initial data, and extracting features according to the initial data to obtain first feature data;

sending a medication reminder to the patient according to the initial data and the prescription data;

deep learning is carried out through a convolutional neural network according to the first characteristic data, and second characteristic data are obtained;

pushing auxiliary treatment information to the patient according to the second characteristic data and the personal information data;

and formulating a patient motivational task based on the second characteristic data and the prescription information data.

2. The method for designing a software for assisting use of a cognitive digital drug according to claim 1, wherein the step of importing doctor's prescription and patient's personal information to obtain prescription data and personal information data comprises:

and transmitting the personal information data and the prescription data from the hospital prescription server to a patient mobile phone end by using a wireless communication mode of a B/S architecture mode, wherein the personal information data and the prescription data are stored by SQLite.

3. The method for designing software for assisting use of a cognitive digital drug according to claim 1, wherein the processing the data information generated by using the digital drug to obtain initial data, performing feature extraction according to the initial data to obtain first feature data comprises:

the method comprises the steps of transmitting medication condition data information of a patient in a wireless communication mode, performing filtering and information model conversion operation through a data conversion interface between a digital medicine and a platform to obtain initial data, and performing feature extraction according to the initial data by adopting a sorting condition mutual information method to obtain first feature data.

4. The method for designing a software for assisting in using a cognitive digital drug according to claim 1, wherein the deep learning is performed through a convolutional neural network according to the first feature data to obtain second feature data, comprising:

establishing a dialogue generating model through a word vector representation method based on a neural network based on the first characteristic data; converting the input first feature data into a computer binary code through a Seq2Seq neural network model introducing an attention mechanism; and performing deep learning based on a convolutional neural network according to the computer binary code to obtain second characteristic data.

5. A method of designing a cognitive digital drug adjuvant software according to claim 1, wherein formulating a patient motivation task based on the second characteristic data comprises:

and acquiring a memory state of the patient in the current test environment, and carrying out dynamic matching of the excitation task according to the mental state of the patient by a learning method of a long-period memory network based on the second characteristic data and the prescription data.

6. A cognitive digital drug coaching software design device for implementing a cognitive digital drug coaching software design method, the device comprising:

the information acquisition module is used for importing doctor prescriptions and personal information of patients to obtain prescription data and personal information data;

the feature extraction module is used for processing data information generated by the digital medicine to obtain initial data, and extracting features according to the initial data to obtain first feature data;

the medicine prompt module is used for sending a medicine prompt to a patient according to the initial data and the prescription data;

the deep learning module is used for performing deep learning through a convolutional neural network according to the first characteristic data to obtain second characteristic data;

the virtual assistant module is used for pushing auxiliary treatment information to the patient according to the second characteristic data and the personal information data;

and the incentive system module is used for formulating a patient incentive task based on the second characteristic data and the prescription information data.

7. The cognitive digital drug coaching software design device according to claim 6, wherein the drug prompt module is further configured to:

and transmitting the personal information data and the prescription data from the hospital prescription server to a patient mobile phone end by using a wireless communication mode of a B/S architecture mode, wherein the personal information data and the prescription data are stored by SQLite.

8. The cognitive digital drug coaching software design device according to claim 6, wherein the drug prompt module is further configured to:

the method comprises the steps of transmitting medication condition data information of a patient in a wireless communication mode, performing filtering and information model conversion operation through a data conversion interface between a digital medicine and a platform to obtain initial data, and performing feature extraction according to the initial data by adopting a sorting condition mutual information method to obtain first feature data.

9. The cognitive digital drug coaching software design device according to claim 6, wherein the virtual helper module is further configured to:

based on the first characteristic data, establishing a dialogue generating model through a word vector representation method based on a neural network, and solving the mapping problem of a cyclic neural network between sequences; converting the input first feature data into a computer binary code through a Seq2Seq neural network model introducing an attention mechanism; and performing deep learning based on a convolutional neural network according to the computer binary code to obtain second characteristic data.

10. The cognitive digital drug coaching software design device according to claim 6, wherein the motivation system module is further configured to:

and acquiring a memory state of the patient in the current test environment, and carrying out dynamic matching of the excitation task according to the mental state of the patient by a learning method of a long-period memory network based on the second characteristic data and the prescription data.

Images