The Research Of H Type Index Based On Multiple Indices

The field of Information Science and The scientific evaluation of the academia have paid a lot of attention to H index since it was brought up by professor Hirsch in 2005.Its theory has been improved and its application has also been broaden. However,this index has some deficiencies because it is calculated only by a single index called cited times. This paper starts from thesis some typical indices such as cited times,download times,author’s h index and periodical’s in the year when thesis is published,annual average on cited times and download times using software SPSS to extract three principle components. The first contains cited times,download times,average on cited times and annual average on download times. The second contains authors’ annual h index. The third one contains periodical annual impact factor. Through analyzing these three comprehensive principle component paper gets the six indices’ weight which are 0.067、0.289、0.157、0.319、0.291、0.258. Paper gets the elected 20 authors’every thesis’total scores according to the six indices’ weight and calculates 20 multiple h type indices Hm in accordance with the method which h index is calculated by. The correlation analysis between Hm and h index tells that these two indices’ linear correlation is significant and the correlation coefficient reaches 0.771 which proves the Hm’s feasibility. Meanwhile,in comparison with h index,Hm which is researched on multiple indices has lower Equal probability,even if different author had the same Hm we also can compare the authors by minimum and precise accurate decimal score. In addition,it’s also useful to improve the unfair phenomenon between the young and old authors to some extent.