Leo Breiman

Timeline

1. 1928–1949

   A physics degree from Caltech

   He is born in New York City in 1928. A physics degree from Caltech follows in 1949.

2. 1954–1968

   Probability theory at UCLA

   He is hired to teach probability theory at UCLA in 1954. He publishes the textbook Probability in 1968.

3. 1968–1980

   Consulting for government and industry

   In 1968 he resigns from UCLA and starts work as a consultant for the US government and industry. He develops statistical methods to predict patterns in everything from traffic to pollution.

4. 1980–1993

   Statistical computing at Berkeley

   In 1980 he joins the UC Berkeley statistics faculty and establishes its Statistical Computing Facility. He retires in 1993 but stays active as a professor in the graduate school.

5. 1984

   Classification and regression trees

   In 1984 Breiman and three co-authors first introduce CART, the classification and regression tree method. The book Classification and Regression Trees is considered his best-known work.

6. 1994–2001

   Bagging and random forests

   In 1994 he develops bagging, which combines classifications from randomly generated training sets. His work on random forests is published in 2001.

7. 2001

   National Academy of Sciences membership

   In May 2001 he is elected to the National Academy of Sciences, one of eight new Berkeley faculty members.

8. 2005

   The SIGKDD Innovations Award

   He wins the 2005 SIGKDD Innovations Award. He dies in 2005 at his Berkeley home after a long battle with cancer.