This week, I am working to get the dynamic time warping functionality into my program. The process of doing so includes re-processing the features to include the time series, putting each series back together when we construct sequences, and then performing the DTW to generate a number that will be used to compute the kNN of each sequence which can then be used for predictions with the models. The processing time of these activities has gone up significantly since we have been using five different metrics with each of the F phase datasets. I am returning to school next week, and once I've completed the DTW processing all that will remain before we put together our second paper (The date for the reach journal we would like to submit it to is October 1), I am hoping I will have time to look again into the Agglomerative Hierarchical Clustering concept, which I did not successfully complete when we explored it earlier in the summer and then changed focus to the paper. We heard back
Before I discuss my work this week, I want to include a few definitions to clarify the language I've been using to describe various metrics.
Day: While a different 'day' period has been defined throughout the course of our research, we have decided to consistently use the 06:00:00 to 20:59:00 period going forward.
Night: Corresponding to 'day', 'night' will be from 21:00:00 to 05:59:00.
Baseline: the first three days of a patient's enrollment in the study. For the sake of Dr. Weeks' experiment, this is a significant period because the patients are wearing Actigraphs but not receiving light therapy. For the sake of our experiment, the significance of these days is slightly different. While we are not primarily concerned with the outcome of light therapy, it is valuable to have a point of comparison for a patient's condition. Utilising multiple days for this purpose allows us to wash out any effect that acclimating to the new watch may have on activity and/or sleep.
Baseline Day Activity Median: This is the median of all activity counts from the first three daytime periods of enrollment. Note that the first daytime period is typically not a full day.
Baseline Night Activity Median: This is the median of all activity counts from the first three nighttime periods of enrollment.
Active Minutes: The number of minutes during a day period where activity counts were greater than the Baseline Day Activity Median.
Inactive Minutes: The number of minutes during a night period where activity counts were were less than or equal to the Baseline Night Activity Median.
This week, I worked on looking at the relationship between my measure of inactivity and Dr. Skornyakov's sleep/wake classification. The reasoning behind my desire to find an alternative to this classification follows the revelation of the difference between the Actigraph sensitivity between the first 26 patients and the rest of the group. Finding an individualized measure for inactivity should, in theory, provide us with a data-driven metric that could allow us to consider patients relative to one another.
My rationale behind using the baseline for this point of reference is explained above, and the median was chosen as the statistic because graphing raw activity data showed that activity counts did not make up a normal distribution and that we had many potential erroneous super-high activity counts that may affect a measure of centrality like the mean.
To measure this relationship, I am going to find the correlation between the calculated number of minutes of sleep per night for each patient with the number of inactive minutes. I will then explore this relationship further by seeing if there is a higher or lower correlation between inactive minutes and the previous day's activity and/or next day's activity.
Day: While a different 'day' period has been defined throughout the course of our research, we have decided to consistently use the 06:00:00 to 20:59:00 period going forward.
Night: Corresponding to 'day', 'night' will be from 21:00:00 to 05:59:00.
Baseline: the first three days of a patient's enrollment in the study. For the sake of Dr. Weeks' experiment, this is a significant period because the patients are wearing Actigraphs but not receiving light therapy. For the sake of our experiment, the significance of these days is slightly different. While we are not primarily concerned with the outcome of light therapy, it is valuable to have a point of comparison for a patient's condition. Utilising multiple days for this purpose allows us to wash out any effect that acclimating to the new watch may have on activity and/or sleep.
Baseline Day Activity Median: This is the median of all activity counts from the first three daytime periods of enrollment. Note that the first daytime period is typically not a full day.
Baseline Night Activity Median: This is the median of all activity counts from the first three nighttime periods of enrollment.
Active Minutes: The number of minutes during a day period where activity counts were greater than the Baseline Day Activity Median.
Inactive Minutes: The number of minutes during a night period where activity counts were were less than or equal to the Baseline Night Activity Median.
This week, I worked on looking at the relationship between my measure of inactivity and Dr. Skornyakov's sleep/wake classification. The reasoning behind my desire to find an alternative to this classification follows the revelation of the difference between the Actigraph sensitivity between the first 26 patients and the rest of the group. Finding an individualized measure for inactivity should, in theory, provide us with a data-driven metric that could allow us to consider patients relative to one another.
My rationale behind using the baseline for this point of reference is explained above, and the median was chosen as the statistic because graphing raw activity data showed that activity counts did not make up a normal distribution and that we had many potential erroneous super-high activity counts that may affect a measure of centrality like the mean.
To measure this relationship, I am going to find the correlation between the calculated number of minutes of sleep per night for each patient with the number of inactive minutes. I will then explore this relationship further by seeing if there is a higher or lower correlation between inactive minutes and the previous day's activity and/or next day's activity.
Comments
Post a Comment