CN106651517B - Drug recommendation method based on hidden semi-Markov model - Google Patents
Drug recommendation method based on hidden semi-Markov modelDownload PDFInfo
- Publication number
- CN106651517B CN106651517BCN201611184351.8ACN201611184351ACN106651517BCN 106651517 BCN106651517 BCN 106651517BCN 201611184351 ACN201611184351 ACN 201611184351ACN 106651517 BCN106651517 BCN 106651517B
- Authority
- CN
- China
- Prior art keywords
- user
- model
- state
- concerned
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Recommending goods or services
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
Translated fromChinese
Description
Technical Field
The invention relates to a medicine recommendation method based on a hidden semi-Markov model.
Background
The medicine polymerization supply chain cooperation platform based on the cloud computing and big data technology is a big data resource and public service cloud platform of a medicine whole industrial chain, has the functions of big data fusion and storage, platform big data mining and application, medicine supervision, industry information integration and the like, integrates the resources of medicine supply chain enterprises, is beneficial to standardizing the economic order of the medicine online trading market, and promotes the healthy and benign development of the medicine whole industrial chain. Under the situation of rapid development of medical big data application services, how to provide accurate medical services for users becomes a key problem which needs to be solved urgently by each big medical platform, and the existing solution generally has the defects of inaccurate correlation of recommendation results and the like.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a medicine recommendation method based on a hidden semi-Markov model. The method comprises the steps of firstly, predicting a disease condition concerned by a user according to an observable network behavior sequence of the user on a cloud platform, and then recommending related medicines to the user according to the disease condition most concerned by the user; the problem of accurate recommendation of user medicines on a medicine polymerization supply chain collaboration platform is solved.
In order to achieve the above object, the present invention provides a drug recommendation method based on hidden semi-markov model, which mainly comprises the following steps:
firstly, training data preprocessing, namely performing data cleaning on the internet behavior data of a user on a medical platform to generate a training data set of a user behavior sequence;
secondly, estimating parameters of a medicine recommendation model;
thirdly, acquiring an internet behavior sequence of a user on a medicine platform;
fourthly, the online behavior sequence of the user is used as an observed value, and the trained medicine recommendation model is used for deducing the state sequence of the user;
fifthly, calculating the expected duration of each state of the state sequence;
sixthly, sequencing the expected duration of each state in a descending order to obtain a plurality of states most concerned by the user, namely a plurality of diseases most concerned by the user;
and seventhly, recommending corresponding medicines to the user according to the previous plurality of diseases most concerned by the user.
Preferably, the drug recommendation model is a hidden semi-markov model based model.
Preferably, the parameter model of the drug recommendation model is represented as: θ ═ pi, a, B }; wherein, pi is the initial state probability of the initial model, A is the state transition probability, and B is the observation probability.
Preferably, the method for estimating the parameters of the drug recommendation model is a forward-backward algorithm.
Preferably, the fourth step of inferring the state sequence of the user using the trained drug recommendation model is based on a Viterbi algorithm.
So-called observation space, which is used for the cooperation of users in the medical aggregation supply chainSequence of online behaviors on the platform, denoted x ═ x1,x2,...,xTThe information processing method includes pages browsed by a user on systems or platforms such as an APP, a medicine cloud platform and a robot, accessed resources or proposed problems and the like. The value space of the state is the disease condition concerned by the user and is expressed as y ═ y1,y2,...yn。
According to the method and the system, the illness state concerned by the user is accurately predicted according to the online behavior of the user on the cloud platform, and then the related medicine is recommended to the user according to the illness state most concerned by the user, so that the relevance of the medicine recommendation result is improved.
Drawings
FIG. 1 is a flow chart of a hidden semi-Markov model-based drug recommendation method according to the present invention.
Detailed Description
The invention is described in detail below with reference to the drawings and specific examples, but the invention is not limited thereto.
The observation value space is an online behavior sequence of a user on a collaborative platform of a medicine polymerization supply chain, and is represented as x ═ x1,x2,...,xTThe information processing method includes pages browsed by a user on systems or platforms such as an APP, a medicine cloud platform and a robot, accessed resources or proposed problems and the like. The value space of the state is the disease condition concerned by the user and is expressed as y ═ y1,y2,...yn。
The parameter model of the drug recommendation model is represented as: θ ═ pi, a, B }; wherein, pi is the initial state probability of the initial model, A is the state transition probability, and B is the observation probability.
For convenience in describing the model, the present invention employs the following notation:
1) t: t + d denotes the time sequence starting from t up to t + d, i.e. t, t + 1.
2)S[t-d+1:t]J represents [ t-d +1, t ═ j]The state over the time interval is j, and neither the state of t +1 nor t-1 is j.
3)St]J denotes that the state at and before time t is j, and the state at t +1 is not j.
4)S[tJ denotes that the state at and after time t is j, and the state of t-1 is not j.
The parameter estimation task of the drug recommendation model is to estimate the corresponding parameters of the hidden semi-Markov model by the collected user online behavior sequence. The invention adopts a forward and backward algorithm to solve the problem of parameter estimation of a medicine recommendation model, which is specifically described as follows.
1) Defining a forward and backward variable:
αt(j,d)＝P[St-d+1:t＝j,o1:t|θ]；
βt(j,d)＝P[ot+1:T|St-d+1:t＝j,θ]。
2) initialization of a forward and backward algorithm:
α1(j,d)＝πj，
βT(j,d)＝1。
3) and (3) iterative derivation process:
4) calculating an intermediate variable:
ηt(j,d)＝P[S[t-d+1:t]＝j,o1:T|λ]＝αt(j,d)βt(j,d)；
ξt(i,d'；j,d)＝P[St]＝i,S[t+1:t+d]＝j,o1:T|λ]＝αt(i,d')a(i,d')(j,d)bj,d(ot+1:t+d)βt+d(j,d)；
5) formula for updating parameters
Wherein when
When the temperature of the water is higher than the set temperature,
otherwise
Given the behavior sequence of the user, the Viterbi algorithm is used for extracting a better user interest sequence y-y1,y2,...yt。
The desired duration of computing state i is:
the total expected time of the user in a certain state may be displayed, i.e. the degree to which the user is interested in a certain disease state may be reflected. Generally, the level of interest a user has in a particular medical condition is proportional to the desired duration of the condition for which the condition is associated.
Referring to fig. 1, an embodiment of the present invention provides a hidden semi-markov model-based drug recommendation method, which mainly includes the following steps:
firstly, training data preprocessing, namely performing data cleaning on the internet behavior data of a user on a medical platform to generate a training data set of a user behavior sequence;
secondly, estimating parameters of a medicine recommendation model;
thirdly, acquiring an internet behavior sequence of a user on a medicine platform;
fourthly, the online behavior sequence of the user is used as an observed value, and the trained medicine recommendation model is used for deducing the state sequence of the user;
fifth, calculating the expected duration of each state of the state sequence
Sixthly, sequencing the expected duration of each state in a descending order to obtain the first N states most concerned by the user, namely the first N illness states most concerned by the user;
and seventhly, recommending corresponding medicines to the user according to the previous plurality of diseases most concerned by the user.
The drug recommendation model is a hidden semi-Markov model-based model. The method for estimating the parameters of the drug recommendation model is a forward and backward algorithm. The fourth step is to use the trained drug recommendation model to infer the state sequence of the user based on the Viterbi algorithm.
According to the embodiment of the invention, the illness state concerned by the user is accurately predicted according to the online behavior of the user on the cloud platform, and then the related medicine is recommended to the user according to the illness state most concerned by the user, so that the relevance of the medicine recommendation result is improved.
The invention has been described in detail, but it is apparent that variations and modifications can be effected by one skilled in the art without departing from the scope of the invention as defined by the appended claims.
Claims (5)
1. A drug recommendation method based on a hidden semi-Markov model is characterized by mainly comprising the following steps:
firstly, preprocessing training data through an observed value space of a user and a value space of a user state, namely, performing data cleaning on internet behavior data of the user on a medical platform to generate a training data set of a user behavior sequence;
secondly, estimating parameters of a medicine recommendation model based on the training data set of the user behavior sequence generated in the first step;
thirdly, acquiring an internet behavior sequence of a target user on a medicine platform;
fourthly, taking the internet behavior sequence of the target user in the third step as an observed value, and deducing the state sequence of the target user by using the medicine recommendation model trained in the second step;
fifthly, calculating the expected duration of each state of the target user state sequence;
sixthly, sequencing the expected duration of each state in a descending order to obtain a plurality of states most concerned by the target user, namely a plurality of diseases most concerned by the target user;
and seventhly, recommending corresponding medicines to the target user according to the previous plurality of diseases most concerned by the target user.
2. The hidden semi-Markov model-based drug recommendation method of claim 1, wherein the drug recommendation model is a hidden semi-Markov model-based model.
3. The hidden semi-markov model-based drug recommendation method of claim 1, wherein the parametric model of the drug recommendation model is represented as: θ ═ pi, a, B }; wherein, pi is the initial state probability of the initial model, A is the state transition probability, and B is the observation probability.
4. The hidden semi-markov model-based drug recommendation method of claim 1, wherein the method for estimating the parameters of the drug recommendation model is a forward-backward algorithm.
5. The hidden semi-markov model-based drug recommendation method of claim 1, wherein the fourth step of inferring the sequence of states of the user using the trained drug recommendation model is based on a Viterbi algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611184351.8ACN106651517B (en) | 2016-12-20 | 2016-12-20 | Drug recommendation method based on hidden semi-Markov model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611184351.8ACN106651517B (en) | 2016-12-20 | 2016-12-20 | Drug recommendation method based on hidden semi-Markov model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106651517A CN106651517A (en) | 2017-05-10 |
| CN106651517Btrue CN106651517B (en) | 2021-11-30 |
Family
ID=58834723
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611184351.8AActiveCN106651517B (en) | 2016-12-20 | 2016-12-20 | Drug recommendation method based on hidden semi-Markov model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106651517B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110085292B (en)* | 2019-04-28 | 2022-07-26 | 广东技术师范大学 | Drug recommendation method, device and computer-readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102270212A (en)* | 2011-04-07 | 2011-12-07 | 浙江工商大学 | User interest feature extraction method based on hidden semi-Markov model |
| CN103345503A (en)* | 2013-07-01 | 2013-10-09 | 杭州万事利丝绸科技有限公司 | Silk product individualized recommendation method based on wavelet network |
| CN104008203A (en)* | 2014-06-17 | 2014-08-27 | 浙江工商大学 | User interest discovering method with ontology situation blended in |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6183737B2 (en)* | 2012-09-27 | 2017-08-23 | 国立大学法人 東京大学 | Optimal route search device |
| US8589164B1 (en)* | 2012-10-18 | 2013-11-19 | Google Inc. | Methods and systems for speech recognition processing using search query information |
| US9978362B2 (en)* | 2014-09-02 | 2018-05-22 | Microsoft Technology Licensing, Llc | Facet recommendations from sentiment-bearing content |
- 2016
- 2016-12-20CNCN201611184351.8Apatent/CN106651517B/enactiveActive
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102270212A (en)* | 2011-04-07 | 2011-12-07 | 浙江工商大学 | User interest feature extraction method based on hidden semi-Markov model |
| CN103345503A (en)* | 2013-07-01 | 2013-10-09 | 杭州万事利丝绸科技有限公司 | Silk product individualized recommendation method based on wavelet network |
| CN104008203A (en)* | 2014-06-17 | 2014-08-27 | 浙江工商大学 | User interest discovering method with ontology situation blended in |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106651517A (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108427708B (en) | Data processing method, data processing apparatus, storage medium, and electronic apparatus | |
| Zaitcev et al. | A deep neural network application for improved prediction of $\text {HbA} _ {\text {1c}} $ in type 1 diabetes | |
| Shany et al. | Are we stumbling in our quest to find the best predictor? Over‐optimism in sensor‐based models for predicting falls in older adults | |
| CN107463701B (en) | Method and device for pushing information stream based on artificial intelligence | |
| JP2005505031A5 (en) | ||
| CN111966904A (en) | Information recommendation method based on multi-user portrait model and related device | |
| WO2023047117A1 (en) | A computer-implemented method, data processing apparatus, and computer program for active learning for computer vision in digital images | |
| CN110516175A (en) | A kind of method, apparatus, equipment and the medium of determining user tag | |
| CN108682457A (en) | Patient long-term prognosis quantitative prediction and intervention system and method | |
| US20150025861A1 (en) | Genetic screening computing systems and methods | |
| Joshi et al. | A decision tree based classification technique for accurate heart disease classification and prediction | |
| CN116051329B (en) | Cognitive diagnosis method, system, device and medium for sparse data scenarios | |
| CN117540336A (en) | Time sequence prediction method and device and electronic equipment | |
| CN102013039A (en) | Gene regulatory network constructing method based on dynamic Bayesian network | |
| CN110717599B (en) | Dissociative representation learning method and device integrating multiple modalities | |
| CN106651517B (en) | Drug recommendation method based on hidden semi-Markov model | |
| Old et al. | Entering the new digital era of intensive care medicine: an overview of interdisciplinary approaches to use artificial intelligence for patients’ benefit | |
| CN106126541B (en) | Personalized solution recommendation method and system based on big data knowledge base | |
| CN120183663A (en) | A personalized health management information recommendation method, system and electronic device | |
| CN109376307A (en) | Article recommended method, device and terminal | |
| CN107786376B (en) | Content pushing method and device and computer equipment | |
| Barkan et al. | Modelling Session Activity with Neural Embedding. | |
| CN112541010A (en) | User gender prediction method based on logistic regression | |
| Chen et al. | From tie strength to function: Home location estimation in social network | |
| CN113742532B (en) | User portrayal method, apparatus and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information | Address after:510665 No. 293 Shipai Zhongshan Avenue, Tianhe District, Guangzhou City, Guangdong Province Applicant after:GUANGDONG POLYTECHNIC NORMAL University Address before:510665 No. 293 Shipai Zhongshan Avenue, Tianhe District, Guangzhou City, Guangdong Province Applicant before:GUANGDONG POLYTECHNIC NORMAL University | |
| GR01 | Patent grant | ||
| GR01 | Patent grant |