Unwanted mail discriminating apparatus and unwanted mail discriminating method

An unwanted mail discriminating apparatus comprising:

• mail receiving means for receiving e-mail;
• information extracting means for extracting discrimination information to be used to determine whether or not the e-mail is unwanted mail from the e-mail received by the mail receiving means, the discrimination information extracted by the information extracting means including sender information specifying a sender of the e-mail;
• database connecting means for connecting to a reliability evaluation database storing information corresponding to the discrimination information, for evaluating reliability of the discrimination information, the reliability evaluation database, to which the database connecting means connects, stores information of a contractual relation between a recipient and the sender of the e-mail;
• reliability evaluating means for evaluating the reliability of the discrimination information extracted by the information extracting means, with reference to the information stored in the reliability evaluation database, to which the database connecting means connects, the reliability evaluating means defining an e-mail group by identifying all e-mail having the same sender information, the reliability evaluating means determining a ratio of a number of e-mails in the e-mail group from the sender of the e-mail wherein exists a contractual relationship between the sender and a corresponding recipient to a total number of e-mails in the e-mail group received by the mail receiving means, the contractual relationship being stored in the reliability evaluation database;
• and
• discriminating means for discriminating whether the e-mail received by the mail receiving means is unwanted mail, based on the reliability of the discrimination information evaluated by the reliability evaluating means, the discriminating means determining the e-mail received by the e-mail receiving means is unwanted mail if the ratio is less than a threshold;
• wherein the reliability evaluating means evaluates for an e-mail group containing identical discrimination information, the reliability of the discrimination information contained in the e-mail group, and wherein the discriminating means discriminates whether the e-mail group is unwanted mail, based on the reliability of the discrimination information evaluated for the e-mail group by the reliability evaluating means;
• wherein the discrimination information extracted by the information extracting means contains link information for access to a site on a communication network, wherein the reliability evaluation database, to which the database connecting means connects, stores information of an access count to the site, for each recipient of the e-mail, and wherein the reliability evaluating means evaluates the reliability, based on a distribution of access counts to the site by recipients of the e-mail in the e-mail group;
• wherein the discriminating means determines the threshold based on the desired safety factor according to the following equation:
• SF=nCx(n!/(n−x)!){(N−n)!/((N−n)(n−x))!}/{N!/(N−n)!},
• where SF is the desired safety factor, nCx is a binomial coefficient, N is a number of users registered in the reliability evaluation database, n is the total number of e-mails in the group, and x/n is the threshold.