基于统计学的Web论坛增量更新策略研究

RESEARCH ON INCREMENTAL UPDATING STRATEGY OF WEB FORUM BASED ON STATISTICS

传统预测网页变化的模型将一种规律应用到所有网页之上,没有考虑各页面之间的区别,针对网络论坛索引页面提出了一种基于统计学规律的增量更新策略模型。通过相关论坛版块的索引页面进行数据的采集,观察并证明其变化大致呈现以日为周期的规律性变化,一日之内的变化曲线与人们的生活规律相吻合。然后采用最小二乘法多项式曲线拟合对其进行数学建模,得到合适的数学模型,并将其应用在索引页面的增量更新之上,从而可以准确预测索引页面下一次更新的时间间隔。实验结果表明,该模型在10%误差范围内,预测的准确率为93.9%。

The traditional model of forecasting page changes applies a rule to all pages, without regard to the differences between pages. In this paper, we propose an incremental updating strategy model based on statistical rules for indexing web pages. Through the data collection and observation of the index page of the relevant forum, it is found that the index page shows a regular change in the daily cycle, and the curve of variation within a day coincides with the law of peopled life. The mathematical model is established by using the least square polynomial curve fitting, and it is applied to incremental updating of the index page, which can predict the time interval of the next updating of the index page. The experimental results show that the accuracy of the model is 93.9% within the 10% error range.