Occam diagnostic method description:

1. Calculation method

Identify the factor that has the most significant impact on data variability, then eliminate that factor. Subsequently, within the remaining data, isolate the factor with the greatest effect on changes, and continue in this manner..

For instance, if it is first determined that "Channel=Google" has the highest impact on data variability, after removing the data for "Channel=Google," it is found that "Country=US" exerts the largest influence on data changes.

2. Operation Approach

Click "+" to expand, indicating that the analysis is performed using the Occam's razor algorithm in step 1 if the current conditions are met.

For example, in step 1, find the two influencing factors "Channel =Google" and "Country =US" :

If you expand on "Channel=Google," it signifies that the Occam's Razor algorithm analysis will be performed filtering by "Channel=Google."

If you expand on "Country=US," it means that the analysis using Occam's Razor algorithm will proceed with the filters set to "non-Google channels and the target country being US."

 

General mode description

1. Calculation Method

Identify the top 10 factors with the most significant data changes, where there may be overlaps between different factors.

For instance: If there is an increase in daily expenses by $2,000, and upon analyzing by channel, it is found that "Channel=Google" has an increase of $1,200, ranking it first; when analyzed by deployment region, "Country=US" sees an increase of $1,000, placing it second. There might be overlap between the data for "Channel=Google" and "Country=US."

 

2. Operation Approach

Clicking "+" to expand indicates that under the current conditions, the search scope is narrowed, and then an analysis is conducted to determine the top 10 factors with the largest data changes.