1. A method comprising:
a) gathering a first time series of event data;
b) gathering a second time series of event data;
c) partitioning a timeline of the first time series of data into at least one time segment based on the second time series event data and a predetermined time window duration;
d) counting a number of events of the second time series in each of the at least one time segment; and
e) determining a likelihood of a dependency relationship of the first time series from the second time series based on the number of events in each of the at least one time segment.
2. The method of claim 1, wherein determining a likelihood of a dependency relationship of the first time series from the second time series includes determining a probability of the first time series given the second time series as a function of a normalization constant.
3. The method of claim 2, further comprising determining a likelihood of no dependency relationship of the first time series including determining a probability of the first time series as a function of a normalization constant.
4. The method of claim 3, further comprising selecting a maximum of the likelihood of a dependency relationship of the first time series from the second time series, and the likelihood of no dependency relationship of the first time series.
5. The method of claim 4, further comprising gathering a third time series of event data and determining a likelihood of a dependency relationship of the first time series from a combination of the second time series and the third time series.
6. The method of claim 2, wherein determining a probability of the first time series given the second time series includes approximating a probability distribution of a weighting parameter.
7. The method of claim 1, wherein determining a likelihood of a dependency relationship includes formulating a generalized linear model for a hazard rate function based on the number of events in each time segment.
8. The method of claim 7, wherein formulating a generalized linear model includes modeling the hazard rate function as an exponential weighting function.
9. The method of claim 4, further comprising determining a dependency structure of a system comprising the first time series and the second time series by determining a likelihood of a dependency relationship for each of the first and second time series of data.
10. The method of claim 1, further comprising predicting an event of the first time series based on the second time series and the determined dependency relationship of the first time series from the second time series.
11. The method of claim 1, further comprising displaying the determined dependency relationship of the first time series on the second time series on a display device of a computer system, wherein the first time series is displayed as a first node and the second time series is displayed as a second node, and the dependency relationship is displayed as a directed connector from the second node to the first node, and a strength of the determined dependency relationship is associated with the directed connector.
12. One or more computer readable media containing computer executable instructions that, when implemented, perform a method comprising:
a) storing a first time series of data as a sequence of time stamps, each time stamp being associated with one of a state up indicator and a state down indicator;
b) storing a second time series of data as a sequence of time stamps, each time stamp being associated with one of a state up indicator and a state down indicator;
c) partitioning a timeline of the first time series of data into at least one time segment based on the second time series data; and
d) determining a likelihood of a dependency relationship of the first time series from the second time series.
13. The one or more computer readable media of claim 12, wherein partitioning includes partitioning only segments of the timeline associated with a state up indicator of the first time series of data.
14. The one or more computer readable media of claim 12, wherein the state up indicator indicates a component up condition and the state down indicator indicates a component down condition.
15. The one or more computer readable media of claim 12, wherein the state up indicator indicates zero events occurring in a predetermined window in the past, and the state down indicator indicates one event occurring in a predetermined window in the past.
16. One or more computer readable media containing executable components comprising:
a) means for selecting a child node from a plurality of time series data, wherein the time series data includes a plurality of events;
b) means for selecting a parent node from the plurality of time series data and a null set, wherein the time series data includes a plurality of events;
c) means for determining a probability of a dependency relationship between the selected child node and the selected parent node including means for counting a number of events of the selected parent node in a predetermined window of time;
d) means for comparing the determined probability of a dependency relationship with a probability of a dependency relationship between the selected child node and another selected parent node.
17. The one or more computer readable media of claim 16, further comprising means for predicting an event of the child node based upon time series data from the selected parent node.
18. The one or more computer readable media of claim 16, further comprising means for displaying the plurality of nodes and an associated dependency structure.
19. The one or more computer readable media of claim 16, wherein the means for comparing includes means for selecting a maximum of the dependency relationship between the selected child node and the selected parent node and the dependency relationship between the selected child node and another parent node, and means for storing the maximal relationship in a dependency structure data store.
20. The one or more computer readable media of claim 16, wherein the means for counting includes determining a hazard rate based upon the number of events of the parent node in a time window extending a predetermined time into the past.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
What we claim is:
1. A thickness-measuring device comprising:
an X-ray tube unit including an X-ray tube;
an X-ray detector unit including an X-ray detector having a sensor; and
a means for driving the X-ray tube unit and the X-ray detector unit back and forth, in synchronism with each other, widthwise of an object whose thickness is measured and which runs through a pathline between the X-ray tube unit and the X-ray detector unit,
a mask being set above the sensor of the X-ray detector to cover it partly.
2. A thickness-measuring device as claimed in claim 1, wherein:
a plurality of pairs of said X-ray tube units and said X-ray detector units are disposed widthwise of the object; and
a mask is set above the sensor of each X-ray detector to cover it partly.
3. A thickness-measuring device as claimed in claim 1, wherein the mask is shaped and disposed so as to expose part of the sensor, whose width is smaller than the whole width of the sensor, to X-rays.
4. A thickness-measuring device as claimed in claim 1, wherein the means for driving the X-ray tube unit and the X-ray detector unit in a pair comprises:
ball-screw rods in a pair which are disposed parallel to each other and on which the X-ray tube unit and the X-ray detector unit travel, respectively;
linear guides in a pair which are disposed along the ball-screw rods, respectively;
servomotors in a pair which rotate the ball-screw rods, respectively; and
a controller which synchronizes the travel of the X-ray tube unit and the travel of the X-ray detector unit.