Technicians at ActivSoftware recently analyzed over two hundred thousand tokens, also referred to as words, in a bayesian spam filter on one of their busier local email servers. They analyzed dozens of items within the data. The most compelling was the spam to ham ratios. Ham is defined as legitimate email sent by legitimate email senders. The compiled list of over 50 words with the highest spam to ham ratio was the result.

Words like 'click' and 'here' don't show up as high, since they are used so often in legitimate email. The ActivSoftware list also reveals that words like 'madam' which is rarely found in legitimate email, while readily found in spam email, had very high ratios. Using this method the team produced a superior list of words found in spam email ordered not just by frequency of usage in spam, but offset by the frequency of usage in legitimate emails. The words are ordered from highest to lowest Spam to Ham ratio.

This study empowered the coders and business employees at ActivSoftware with an increased ability to continue to track and, according to the team at ActivSoftware, eliminate the negative impacts of spam on the business community. The team at ActivSoftware continues to attempt to learn spammers in order to increase the effectiveness and deliverability for customers using XM Mail Server.

Here's the top ten:

homeowner discreet madam materially unclaimed anticipates soma preapproved unconditionally beneficiary

Note: This list has been updated and might change frequently as we continue to hone it in. The newer version, 8/8/05, also includes bid information from Overture as well as the Spam:Ham ratios.

See the newest spam to ham word list

Technicians at ActivSoftware, recently analyzed over two hundred thousand tokens, also referred to as words, in a bayesian spam filter on one of their busier local email servers. They analyzed dozens of items within the data. The most compelling was the spam to ham ratios. Ham is defined as legitimate email sent by legitimate email senders. The compiled list of over 50 words with the highest spam to ham ratio was the result.

Words like 'click' and 'here' don't show up as high, since they are used so often in legitimate email. The list also reveals that words like 'madam' which is rarely found in legitimate email, while readily found in spam email, had very high ratios. Using this method the team produced a superior list of words found in spam email ordered not just by frequency of usage in spam, but offset by the frequency of usage in legitimate emails. The words are ordered from highest to lowest Spam to Ham ratio.

This study empowered the coders and business employees at ActivSoftware with an increased ability to continue to track and, according to the team at ActivSoftware, eliminate the negative impacts of spam on the business community. The team at ActivSoftware continues to attempt to learn spammers in order to increase the effectiveness and deliverability for customers using XM Mail Server.

Here's the top ten:

homeownerdiscreet madam materially unclaimed anticipates soma preapproved unconditionallybeneficiary

Note: This list has been updated and might change frequently as we continue to hone it in. The newer version, 8/8/05, also includes bid information from Overture as well as the Spam:Ham ratios.

See the newest spam to ham word list: http://blog.activsoftware.com/entry/36/top_50_spam_words