Actually, the base unit for mass IS the kilogram, at least in the modern MKS version of the metric system, adopted in 1954. The older CGS system did use grams as the base unit for
mass, but the base unit for length in CGS is the centimeter. There has never been a metric system where both meters and grams were base units. A reference can be found here .
I implemented bayesian filter that uses a database of consisting of both word pairs and single words. The performance of this filter alone is slightly better than that of SpamAssassin, but after noticing that the bayesian filter catches almost all the spam that SpamAsassin misses and vice versa, I decided to try running them in series.
If both SpamAssassin and the bayesian filter agree that a message is spam, it gets routed to my Spam mailbox. If both agree that the message is not spam, it gets delivered to my inbox. In case of a disagreement, the message is stored in a separate mailbox/database where I can manually check it (previously all messages flagged by SpamAssassin went here).
After running the combined filter for a week, the results are quite impressive; Zero false positives, zero false negatives and the amount of messages that I have to check manually has decreased to 1/10 of the previous number.
Slashdot Mirror
Not a typo. Vulture Funds specialize in 'distressed' investments. A money-burning operation like Air Canada certainly qualifies.
Actually, the base unit for mass IS the kilogram, at least in the modern MKS version of the metric system, adopted in 1954. The older CGS system did use grams as the base unit for mass, but the base unit for length in CGS is the centimeter. There has never been a metric system where both meters and grams were base units. A reference can be found here .
I implemented bayesian filter that uses a database of consisting of both word pairs and single words. The performance of this filter alone is slightly better than that of SpamAssassin, but after noticing that the bayesian filter catches almost all the spam that SpamAsassin misses and vice versa, I decided to try running them in series.
If both SpamAssassin and the bayesian filter agree that a message is spam, it gets routed to my Spam mailbox. If both agree that the message is not spam, it gets delivered to my inbox. In case of a disagreement, the message is stored in a separate mailbox/database where I can manually check it (previously all messages flagged by SpamAssassin went here).
After running the combined filter for a week, the results are quite impressive; Zero false positives, zero false negatives and the amount of messages that I have to check manually has decreased to 1/10 of the previous number.