Research And Implementation Of Woodland Suitability Prediction System Based On C4.5 Improvement

With the continuous change of information technology,the level of landscaping is being improved in urban construction.Since the eighteen Party's Congress,landscaping has become a new environmental industry,and its development has been continuously expanding.In combination with the current situation of urban development and the improvement of people's quality of life,the urban garden construction has high requirements on the species,quality and survival rate of the seedlings,especially in the resources and design ability of the seedlings,which will become the core competitiveness of the dominant enterprises.In order to promote the development of the new industry of urban landscape greening,this theme,taking one of the main raw materials of the urban landscape greening project as the main factor,studies the decision tree algorithm in the data mining technology,puts forward an improved C4.5 decision tree algorithm,and formulates the forestland suitability prediction system.First,by analyzing the distribution characteristics of the different values of the continuous attributes under the same attribute,the principle of normal distribution in statistics is combined with the C4.5 algorithm,and the continuous attributes are dispersed on the basis of the characteristics of normal distribution.The segmentation point is set as an unknown number,and the integral number of the normal distribution characteristic function is used to calculate the number of unknown points.Then,the segmentation points are sorted and combined with the Fayyad boundary value principle,the attribute values at the two adjacent boundary points are selected as the test attribute values to calculate the information entropy,and the forest suitability prediction model is established.Finally,we use the program to realize the C4.5 algorithm and other classification algorithms before and after the improvement and compared with the experimental results.The theoretical analysis and experimental results show that the improved C4.5algorithm improves the accuracy of the classification of continuous attributes and the construction speed of the decision tree to a certain extent,and reduces the runningtime of the algorithm.And then combined with data visualization technology,the visual conversion of the planting area and economic benefits of the garden seedlings in recent years is visualized,which provides a powerful support for the government in the decision-making of garden construction,and helps the government to develop the new industry of garden greening.In order to build an applicable forestland suitability prediction system,this theme first studies the composition of the forecast system,carries out functional requirement analysis and non functional requirement analysis to the whole system,then makes a summary design and detailed design of the database,and constructs the decision tree model and prediction model using the improved C4.5 algorithm.Finally,Java development language,B/S architecture model,Spring,Spring MVC,Hibernate framework are used to realize the forest suitability prediction system,a number of users are invited to test the system.It proves the stability of the system and provides comprehensive analysis and intelligence support for the government managers.