Evaluation Model Based On Linear Regression Analysis Of Network Information Resources

Internet is an open global distributed network. It has already developed into a large distributed information space, including such various forms as science, culture, commerce, amusement and so on. Now Internet is a valuable information network. But its opening, distributing, freedom and sharing make internet to be a extensive source information network. Because a lot of random information issue, as Internet offers information to user, it also cause the over-load information and polluted. We need essential to the information resources of the network. It contributes to get the useful information quick and accurately. In fact, the evaluation to the information resources of the network is the evaluation to the standard and method of information quality, which in the information resources of the network should be use reasonable.The model that this text does the best to analyze the information resource of the network in terms of quantity goes on more deeply and completely. At first, it summarizes on the basis of the others' research, chooses appropriate evaluation criteria, sets up a comparatively scientific evaluation criteria system and classifies the criteria to make it thinning. Secondly, based on the system, evaluation model has set up by utilizing statistical linear regression model method. Then, we begin to investigate and get the data, use the data to estimate the parameters in model. So we can describe in detail the method and stop that model use at the application in the model, and it becomes possible to use in reality. Finally, the model that has set up can be used to evaluate the concrete websites, in order to examine the rationality of the model. Using linear regression model statistical method, we can get a overall target of quantization while evaluating. It is good to compare a lot of the criteria, so the inaccurate qualitative evaluation can be avoided. Because the target investigated is the wide user and needn't the expert, the model can hare extensive range of the application. Because it is uses the average of each score to determine the comprehensive scores, the surveyor's individual fancy can be eliminated through the average, then the result can get the objective evaluation to the appraisal object. The model also can weigh the confidence level of the comprehensive scores with confidence interval form, besides the comprehensive scores to the appraisal object, which can let the result be more reliable. At the same time, it has overcome the shortage in the old model, which can only give a comprehensive evaluation, but not to consider the confidence level of the evaluation. This is an important innovation.