Signal processing for making predictive determinations
In some examples, unstructured data is evaluated using a natural language processing model to output a set of subjective indicators. These subjective indicators are scored using a predictive model to determine whether a dependent user has or is likely to develop a particular condition such as a cellular abnormality.
1. A system comprising:
memory configured to store processor-executable instructions; and
one or more processing devices in communication with the memory and configured to execute the processor-executable instructions to perform operations comprising:
monitoring a data stream comprising a plurality of messages originating from a plurality of sending systems in a computer network and transmitted to one or more destination devices for storage;
identifying a particular message from among the plurality of messages in the data stream based on a message type associated with the particular message, the particular message including a note that includes an observation corresponding to a dependent user as observed and recorded by an authorized user, at least a portion of the observation represented as unstructured data in the note;
evaluating the unstructured data in the note using a predictive model to identify subjective indicators present in the note, the subjective indicators indicating an assessment, a plan, or sentiment of the authorized user with respect to the dependent user and a cellular abnormality, the cellular abnormality being a cellular condition potentially present in the dependent user;
assigning a weight value to each of the identified subjective indicators based on a respective correspondence of the identified subjective indicators to the cellular abnormality;
determining a composite abnormality score based on weight values of the identified subjective indicators and relationships between the identified subjective indicators, the composite abnormality score corresponding to a presence of the cellular abnormality in the dependent user;
determining whether the cellular abnormality is present in the dependent user based on a first comparison of the composite abnormality score with a normality threshold;
generating a cellular abnormality report that identifies at least the presence of the cellular abnormality and the dependent user; and
causing transmission the cellular abnormality report to an electronic device and/or user interface mapped to at least one recipient of a set of one or more recipients.
2. The system as recited in claim 1 , wherein the predictive model corresponds to a natural language processing model.
3. The system as recited in claim 2 , wherein the natural language processing model is trained based at least in part on a corpus of the unstructured data mapped to a condition.
4. The system as recited in claim 3 , wherein the evaluating the unstructured data, using the predictive model comprising using the natural language processing model to identify the subjective indicators present in the unstructured data.
5. The system as recited in claim 4 , further comprising generating a subjective assessment score based at least in part on the identifying of the one or more subjective indicators present in the unstructured data, wherein the subjective assessment score indicates a likelihood of the cellular condition occurring.
6. The system as recited in claim 5 , wherein the subjective assessment score corresponds to a period of time in which the condition is indicated as likely to occur.
7. The system of claim 1 , wherein the note comprises a report that describes causes and effects associated with cellular abnormalities.
8. The system of claim 1 , wherein the operations comprise an initial step of maintaining a list of message types that include one or more of notes or other information relevant to making a determination about the cellular abnormality.
9. The system of claim 1 , wherein the subjective indicators include one or more of words, terms, or phrases that are indicative of cellular abnormalities.
10. The system of claim 1 , wherein the normality threshold is a specific to certain classes of dependent users.
11. The system of claim 1 , wherein determining whether the cellular abnormality is present includes determining, based on the weight values of the identified subjective indicators and the relationships between the identified subjective indicators, a plurality of type scores corresponding to a plurality of cellular abnormality types to which the cellular abnormality potentially belongs.
12. The system of claim 11 , wherein the operations further include determining, based on a second comparison of the plurality of type scores with respective type thresholds, that the cellular abnormality present in the dependent user is a particular type of cellular abnormality of the plurality of cellular abnormality types.
13. A method comprising:
monitoring a data stream comprising a plurality of messages originating from a plurality of sending systems in a computer network and transmitted to one or more destination devices for storage;
identifying a particular message from among the plurality of messages in the data stream based on a message type associated with the particular message, the particular message including a note that includes an observation corresponding to a dependent user as observed and recorded by an authorized user, at least a portion of the observation represented as unstructured data in the note;
evaluating the unstructured data in the note using a predictive model to identify subjective indicators present in the note, the subjective indicators indicating an assessment, a plan, or sentiment of the authorized user with respect to the dependent user and a cellular abnormality, the cellular abnormality being a cellular condition potentially present in the dependent user;
assigning a weight value to each of the identified subjective indicators based on a respective correspondence of the identified subjective indicators to the cellular abnormality;
determining a composite abnormality score based on weight values of the identified subjective indicators and relationships between the identified subjective indicators, the composite abnormality score corresponding to a presence of the cellular abnormality in the dependent user;
generating a cellular abnormality report that identifies at least the presence of the cellular abnormality and the dependent user; and
causing transmission the cellular abnormality report to an electronic device and/or user interface mapped to at least one recipient of a set of one or more recipients.
14. The method as recited in claim 13 , wherein the predictive model corresponds to a natural language processing model.
15. The method as recited in claim 14 , wherein the natural language processing model is trained based at least in part on a corpus of the unstructured data mapped to a condition.
16. The method as recited in claim 15 , wherein evaluating the unstructured data, using the predictive model comprising using the natural language processing model to identify the subjective indicators present in the unstructured data.
17. The method as recited in claim 16 , further comprising generating a subjective assessment score based at least in part on the identifying of the one or more subjective indicators present in the unstructured data, wherein the subjective assessment score indicates a likelihood of the cellular condition occurring.
18. The method as recited in claim 17 , wherein the subjective assessment score corresponds to a period of time in which the condition is indicated as likely to occur.
19. One or more non-transitory, machine-readable media storing machine-executable instructions that, when executed by one or more processing devices, cause the one or more processing devices to perform operations comprising:
monitoring a data stream comprising a plurality of messages originating from a plurality of sending systems in a computer network and transmitted to one or more destination devices for storage;
identifying a particular message from among the plurality of messages in the data stream based on a message type associated with the particular message, the particular message including a note that includes an observation corresponding to a dependent user as observed and recorded by an authorized user, at least a portion of the observation represented as unstructured data in the note;
evaluating the unstructured data in the note using a predictive model to identify subjective indicators present in the note, the subjective indicators indicating an assessment, a plan, or sentiment of the authorized user with respect to the dependent user and a cellular abnormality, the cellular abnormality being a cellular condition potentially present in the dependent user;
assigning a weight value to each of the identified subjective indicators based on a respective correspondence of the identified subjective indicators to the cellular abnormality;
determining a composite abnormality score based on weight values of the identified subjective indicators and relationships between the identified subjective indicators, the composite abnormality score corresponding to a presence of the cellular abnormality in the dependent user;
generating a cellular abnormality report that identifies at least the presence of the cellular abnormality and the dependent user; and
causing transmission the cellular abnormality report to an electronic device and/or user interface mapped to at least one recipient of a set of one or more recipients.