Prediction method and system for popularity values of different cities and model training method and system

A technology of model training and popularity value, applied in the field of computer information, can solve the problem that the popularity of different cities cannot be effectively predicted, and achieve the effect of good housing status and inventory optimization

Pending Publication Date: 2020-05-08
CTRIP COMP TECH SHANGHAI
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to overcome the fact that the popularity of different cities cannot be effectively predicted in the prior art, so that service agencies such as hotels cannot make targeted adjustments to the products or services they provide, which affects service quality and user experience. Defects, providing a model training method for travel data prediction, a model training system, a prediction method for heat values ​​in different cities, a prediction system, electronic equipment, and a storage medium

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Prediction method and system for popularity values of different cities and model training method and system
  • Prediction method and system for popularity values of different cities and model training method and system
  • Prediction method and system for popularity values of different cities and model training method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0077] This embodiment provides a model training method for travel data prediction, figure 1 A flowchart of the model training method is shown, including:

[0078] Step 101. Obtain historical travel data about multiple target cities from the OTA platform.

[0079] Step 102, extract feature data from the historical travel data.

[0080] Step 103. Divide the feature data into a training set and a prediction set.

[0081] Step 104, input the feature data in the training set into the deep machine learning model for training.

[0082] Step 105, input the feature data in the prediction set into the trained deep machine learning model for prediction to obtain predicted travel data.

[0083] Step 106 , judging whether the relative error between the predicted travel data and the real travel data is smaller than the error threshold, if yes, execute step 107 , if not, return to step 104 .

[0084] Step 107, confirming that the trained deep machine learning model is a travel data pred...

Embodiment 2

[0090] This embodiment provides a model training method for travel data prediction. This embodiment is an improvement to Embodiment 1, wherein, in order to fully learn the training set in the deep machine learning model, effective features can be extracted. , and can also overcome the situation of overfitting and data leakage. The deep machine learning model in this embodiment includes two layers of machine learners, and the output of the first layer of machine learners is used as the input of the second layer of machine learners. Among them, the first layer of machine learning includes several models so that the data set can be divided and cyclically trained to reduce the risk of overfitting. For example, 5-fold cross-validation can be used, including Xgboost model, LightGBM model, BayesianRidge model, DecisionTreeRegressor model and The RandomForestRegressor model, the second layer machine learner includes five models of the Xgboost model. Among them, since the Xgboost model...

Embodiment 3

[0096] This embodiment provides a method for predicting heat values ​​of different cities, image 3 A flowchart of the forecasting method is shown, including:

[0097] Step 301: Input the city to be predicted into the travel data prediction model to obtain the daily travel data of the city to be predicted within the preset range of future dates.

[0098] Step 302: Input the travel data into a preset mapping function to obtain the popularity level of the city to be predicted.

[0099] Wherein, the travel data prediction model is trained according to the training method of embodiment 1 or embodiment 2. In this embodiment, the trained travel data prediction model can be deployed to online scheduling for predicting the cities that need to be predicted, and When prediction is required, the travel data prediction model can be invoked by loading an online model file.

[0100] Wherein, in one implementation manner of this embodiment, the predicted travel data is arrival number data,...

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

No PUM Login to view more

Abstract

The invention discloses a model training method and a model training system for travel data prediction, a prediction method and a prediction system for popularity values of different cities, electronic equipment and a storage medium. The model training method comprises the steps of obtaining historical travel data about a plurality of target cities; extracting feature data; dividing the feature data into a training set and a prediction set; inputting the feature data in the training set into a deep machine learning model for training; predicting the feature data in the prediction set to obtainpredicted travel data; and judging whether the relative error between the predicted travel data and the real travel data is less than an error threshold, and if so, taking the trained deep machine learning model as a travel data prediction model. According to the method, historical travel data is utilized to train the deep machine learning model to obtain the travel data prediction model, so thatfuture travel data prediction models of different cities can be predicted, and a judgment basis is further provided for prediction of market popularity.

Description

technical field [0001] The present invention relates to the field of computer information technology, in particular to a model training method for travel data prediction, a model training system, a prediction method for heat values ​​of different cities, a prediction system, electronic equipment and a storage medium. Background technique [0002] The OTA (Online Travel Agency) platform provides a large number of data services on the basis of enterprise-level information technology and data advantages represented by mobile Internet, cloud computing, big data, artificial intelligence, Internet of Things and blockchain. [0003] The popularity of different cities is mainly reflected in the changes in the number of travelers (including planes and trains) and the number of hotel room nights. The popularity of different cities shows long-term seasonality, trends, and holiday peaks. Since small-scale hotel merchants (such as economy, comfort, three-star) do not have corresponding i...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q30/02G06Q50/12G06N20/20G06N20/00G06N7/00G06N5/00
CPCG06Q30/0202G06Q50/12G06N20/00G06N20/20G06N5/01G06N7/01
Inventor 黎建辉柳影波郭海山
Owner CTRIP COMP TECH SHANGHAI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap