The invention relates to a detection and
analysis method for the abnormal behaviors of a user in a
big data environment. The method is characterized in that the method comprises the following steps: enabling a user abnormal behavior detection
system to carry out the
abnormality analysis of user access behaviors in an offline mode through
machine learning according to the log
record of the user in HDFS in one historical statistical period, and building a user behavior model; enabling the user abnormal behavior detection
system to carry out the online comparison of real-time behaviors and historical behaviors based on the current real-time user's operation behavior in
Storm; transmitting safety early-warning information to Kaffka and displaying the safety early-warning information at a
Stream interface if the difference between real-time behaviors and historical behaviors is big, or else judging that the behavior is a legal safe behavior. Compared with the prior art, the method supports the definition of a behavior mode or a user portrait according to the historical use behavior
habit of the user at a Hadoop platform through a
machine learning
algorithm. A training
system updates a model each month in a default manner, and the
granularity of the model is one minute.