4 ms·
This might be a hard problem. To illustrate, I'm assuming that you are using the degree of mobility as a proxy for the normal->depressed spectrum. Consider the
by powerbroker 3y ago
This might be a hard problem. To illustrate, I'm assuming that you are using the degree of mobility as a proxy for the normal->depressed spectrum. Consider the following hypothetical:
1. User owns a robust dog that likes walks and rough-housing;
2. User's dog dies;
3. User replaces dog with a puppy, that moves about 1/3 the speed of the first dog.
From 1 to 3, the user's mobility declines sharply. Does your app diagnose depression? I'm sure this hypothetical is the tip of the iceberg. Now you may have other metric you measure, as well. But if gross mobility is the chief input to your algorithm, you may find that many, many more inputs are necessary to even start to approximate a professional's diagnosis.