Hi I am a college student in wireless technology but am only very sketchy on statistics. I'm am trying to write a program to tell me if activity is unusual. I can read the activity of a person from a motion detector from day to day i.e.
mon activity - 250 times
tues - 300
wed - 400
thurs - 275
fri - 300
sat - 340
sun - 440
I've no problem getting mean = 329.3, stan dev = 68.7, median = 300
What i want to be able to do is find some algorithm to say that on day 8 if there is no activity (0) or very low that it is unusual or some confidence level. can anyone tell me any possiblilites or where I should be looking?
To determine if the activity on day 8 is unusual or falls outside a specific confidence level, you can use a statistical technique called z-score.
The z-score measures how many standard deviations a particular value is from the mean. By calculating the z-score for day 8, you can determine whether the activity on that day is significantly different from the average activity.
Here's how you can calculate the z-score:
1. Subtract the mean from the activity value of day 8.
z = (activity_day8 - mean) / standard_deviation
2. Compare the calculated z-score to a critical value or threshold. The critical value determines the level of confidence you want to set. A commonly used critical value is 1.96, which corresponds to a 95% confidence level. If the calculated z-score is greater than or less than the critical value, then the activity on day 8 can be considered unusual.
For example, let's say the activity on day 8 is 0. Plug in the values to calculate the z-score:
z = (0 - 329.3) / 68.7
The calculated z-score for day 8 would be -4.79.
Since -4.79 is significantly smaller than -1.96, you can conclude with a high level of confidence that the activity on day 8 (0) is unusual.
To implement this in your program, you can define a threshold z-score and compare the calculated z-score to that threshold. If the z-score exceeds the threshold, you can notify that the activity is unusual.