语言处理常用算法小论文.docVIP

Authorship analysis by means of word class in Tom Sawyer and Oliver Twist Introduction First of all, we can show the outline of this report simply. We make full use of different computational software to deal with the two novels and access the various pos values of each novel. By analyzing these values or proportions, we can reach some suggestions or conclusions about different authors. The meaning to do so is to provide actual evidence to existing assumptions or hypotheses. And the software used in this process contains antcon, search and replace, SPSS and visual foxpro. Methodology We will show the details in the following. There are two ways to generate the pos proportion of the two files. Firstly, we can make use of the visual foxpro to output the pos result directly. The detailed programs are attached to the end of this report. You can access the pos data by this method in one step. It is not difficult to do so for students who handle visual foxpro with great skill. But for other ones they can turn to the second way. There are two softwares to introduce in this method – search and replace, and antconc. We search the “space” in each file, and then replace it with “hard return” (CHR-13). Then we can open this new file in antcon. By the wordlist column, we can save the output to text file. In this antconc result file, we can get the number of each part of speech. The ways to analyze the pos result files. We all know that all ways lead to Rome. There are also several ways to deal with the pos data. The most common one is to use excel of office – import external data. After we change the text file to excel file, we can make full use of the auto filter function in excel to filter different kinds of pos. Here the pos code file is necessary. We classify the same word class together and calculate the total number of it. For example, we choose the user-defined filter function to display part of speech starting with N (N*), we can conclude that the following word clas