Policy text classification method based on machine learning
A text classification and machine learning technology, applied in machine learning, neural learning methods, instruments, etc., can solve problems such as unguaranteed accuracy and inapplicability to changing scenarios, and achieve the effect of ensuring accuracy and intelligent policy classification
Pending Publication Date: 2021-04-16
广州博士信息技术研究院有限公司
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Hard-coded scripts cannot be applied to changing scenarios, and the accuracy cannot be guaranteed
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0056] Such as figure 1 As shown, the present embodiment provides a policy text classification method based on machine learning, and the method includes the following steps:
[0057] S1), label the prepared data according to policy alerts, policies and regulations, declaration notices, policy announcements, policy activities, policy interpretations, and single-text labels;
[0058] S2), using the data of each category as a training data set, and randomly dividing it into 80% training set, 15% test set and 5% evaluation set;
[0059] S3), giving the title of the policy a weight of 0.7 and a content weight of 0.3, merging them into text, and converting the text into a vector;
[0060] S4), choose the CNN algorithm as the basic algorithm, add policy dimension function, weight factor, configure 100 layers, batch 100, learning rate 2e-5, model save interval 500 input word vector sequence, generate a feature map, use The maximum pooling operation in the time dimension obtains the ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a policy text classification method based on machine learning. The policy text classification method comprises the following steps: S1) carrying out classification annotation and single text annotation on data; s2) taking each category of data as a training data set, S3) combining titles of policies into a text, and converting the text into a vector; s4) selecting a CNN algorithm as a basic algorithm, splicing the features obtained by all convolution kernels to obtain fixed-length vector representation of the text, and connecting the text classification problem to softmax to construct a complete model; s5) defining output data; s6) adding the vector into the tuple; s7) defining a classifier; s8) defining a loss function and an accuracy function; S9) defining an optimization method; S10) performing network training; s11) carrying out model evaluation; and S12) carrying out model prediction. The method has an efficient, accurate and intelligent policy classification effect, and the problem of large-scale processing of policy data classification is solved.
Description
technical field [0001] The invention relates to the technical field of policy classification, in particular to a policy text classification method based on machine learning. Background technique [0002] Social Policy and Human Development The new perspective of human development emphasizes a human-centered development concept, emphasizing that economic and social development should revolve around meeting various needs of people, improving people's quality of life, and solving various problems faced by human society. [0003] Industrial policy, as an auxiliary tool used by the country to develop industries, plays an important role in optimizing the industrial structure and can promote the sound and rapid development of the national economy. [0004] In recent years, the central and local governments have successively issued various policies to protect people's livelihood and promote the economy. Especially in response to the impact of the epidemic, various localities and dep...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F40/289G06F40/242G06K9/62G06N3/04G06N3/08G06N20/00G06Q50/26
Inventor 龚增辉胡建敏
Owner 广州博士信息技术研究院有限公司