Systems and methods for classifying data in building automation systems
A computerized method of assigning a building automation system point type to a plurality of unclassified data points is provided. The method includes receiving unclassified data points and attributes for each data point. The method includes receiving classifications for a first subset of the unclassified data points. Each classification associates a data point with a building automation system point type. The method includes generating a term set containing substrings that appear in the attributes. The method includes generating a first matrix describing a frequency that the substrings appear in the attributes. The method includes calculating an indicator of a probability that the presence of the selected substring results in the data point belonging to the selected point type. The method includes assigning a point type to a second subset by finding the substring and potential point type pair having the greatest indication of probability.
1. A building automation system (BAS) comprising:
a plurality of BAS devices comprising sensors and building equipment configured to monitor and control one or more building spaces using a plurality of data points, wherein a number of the plurality of BAS devices are associated with one or more unclassified data points of the plurality of data points, each of the one or more unclassified data points comprising an attribute containing a substring, the substring comprising one or more characters forming a text description of the attribute; and
a processing circuit configured to generate a set of probability values for each unclassified data point, each probability value of the set of probability values corresponding to a different BAS point type and indicating a probability that the unclassified data point belongs to the corresponding BAS point type based on the one or more characters of the substring contained in the attribute of the unclassified data point and a classification technique, wherein the one or more characters are an input to the classification technique and the set of probability values are an output of the classification technique;
wherein the processing circuit is configured to assign one BAS point type to each of the one or more unclassified data points by identifying the BAS point type corresponding to a greatest probability value in each set of probability values.
2. The building automation system of claim 1 , wherein the processing circuit comprises a memory device storing a list of known BAS point types;
wherein the processing circuit is configured to assign one known BAS point type of the list of known BAS point types to each of the one or more unclassified data points.
3. The building automation system of claim 1 , wherein the processing circuit comprises a word breaker configured to:
identify a plurality of classified data points having known BAS point types, each of the plurality of classified data points comprising an attribute containing a substring; and
generate a term set comprising a plurality of substrings contained in the attributes of the plurality of classified data points.
4. The building automation system of claim 3 , wherein the processing circuit comprises a term frequency calculator configured to calculate, for each of the plurality of substrings in the term set, a frequency value indicating a frequency that the substring appears in the attributes of the plurality of classified data points.
5. The building automation system of claim 4 , wherein the processing circuit is configured to generate the set of probability values for each unclassified data point based on the calculated frequency values.
6. The building automation system of claim 1 , wherein the classification technique comprises a naive Bayes classification technique, wherein the processing circuit is configured to generate the set of probability values for each unclassified data point using the naive Bayes classification technique.
7. The building automation system of claim 1 , wherein the classification technique comprises latent semantic indexing, wherein the processing circuit is configured to generate the set of probability values for each unclassified data point using the latent semantic indexing.
8. A building automation system (BAS) comprising:
a plurality of BAS devices comprising sensors and building equipment configured to monitor and control one or more building spaces with a plurality of data points, wherein each of the plurality of BAS devices is associated with one or more unclassified data points of the plurality of data points, each of the one or more unclassified data points comprising an attribute containing a substring, the substring comprising one or more characters forming a text description of the attribute; and
a processing circuit configured to identify a set of classified data points having known BAS point types, the processing circuit comprising:
a memory device storing a list of the known BAS point types and a term set comprising a plurality of substrings contained in attributes of the set of classified data points;
a learner configured to determine, for each combination of a plurality of combinations of a substring selected from the plurality of substrings in the term set and a BAS point type selected from the list of the known BAS point types, a probability that each unclassified data point belongs to the selected BAS point type based on the one or more characters of the substring contained in the attribute of the unclassified data point and a classification technique, wherein the one or more characters are an input to the classification technique and the probability is an output of the classification technique; and
a classifier configured to assign one BAS point type from the list of the known BAS point types to each of the one or more unclassified data points by identifying the combination of the selected substring and the selected BAS point type with a greatest probability.
9. The building automation system of claim 8 , wherein the processing circuit comprises a word breaker configured to generate the term set comprising the plurality of substrings contained in the attributes of the set of classified data points.
10. The building automation system of claim 8 , wherein the processing circuit comprises a term frequency calculator configured to calculate, for each of the plurality of substrings in the term set, a frequency value indicating a frequency that the substring appears in the attributes of the set of classified data points.
11. The building automation system of claim 10 , wherein the learner is configured to determine the probability for each combination of the selected substring and the selected BAS point type based on the calculated frequency values.
12. The building automation system of claim 8 , wherein the classification technique comprises a naive Bayes classification technique, wherein the learner is configured to determine the probability for each combination of the selected sub string and the selected BAS point type using the naive Bayes classification technique.
13. The building automation system of claim 8 , wherein the classification technique comprises latent semantic indexing, wherein the learner is configured to determine the probability for each combination of the selected substring and the selected BAS point type using the latent semantic indexing.
14. A building automation system (BAS) comprising:
a plurality of BAS devices comprising sensors and building equipment configured to monitor and control one or more building spaces with a plurality of data points, wherein each of a number of the plurality of BAS devices is associated with one or more unclassified data points of the plurality of data points, each of the one or more unclassified data points having a plurality of attributes comprising a first attribute containing a substring and a second attribute containing a numeric value, the substring comprising one or more characters forming a text description of the first attribute; and
a processing circuit comprising:
a memory device storing a list of substrings and a list of BAS point types; and
a classifier configured to filter at least one of the list of substrings and the list of BAS point types based on the plurality of attributes of the one or more unclassified data points;
wherein the classifier is configured to assign one of the BAS point types to each unclassified data point based on the one or more characters of the substring contained in the first attribute of the unclassified data point, the numeric value contained in the second attribute of the unclassified data point, and a classification technique, wherein the one or more characters are an input to the classification technique and a set of probability values are an output of the classification technique.
15. The building automation system of claim 14 , wherein the memory device stores a skip list identifying one or more substrings in the list of substrings;
wherein the classifier is configured to filter the list of substrings to remove the one or more substrings identified by the skip list.
16. The building automation system of claim 14 , wherein the processing circuit comprises a word breaker configured to:
identify a plurality of classified data points having known BAS point types, each of the plurality of classified data points comprising an attribute containing a substring; and
generate a term set comprising a plurality of substrings contained in the attributes of the plurality of classified data points.
17. The building automation system of claim 16 , wherein the processing circuit comprises a term frequency calculator configured to calculate, for each of the plurality of substrings in the term set, a frequency value indicating a frequency that the substring appears in the attributes of the plurality of classified data points.
18. The building automation system of claim 14 , wherein the processing circuit comprises a learner configured to generate the set of probability values for each unclassified data point with the classification technique, each probability value of the set of probability values corresponding to a different BAS point type and indicating a probability that the unclassified data point belongs to the corresponding BAS point type based on the one or more characters of the substring contained in the first attribute of the unclassified data point and the classification technique.
19. The building automation system of claim 18 , wherein the classification technique comprises a naive Bayes classification technique, wherein the learner is configured to generate the set of probability values for each unclassified data point using the naive Bayes classification technique.
20. The building automation system of claim 18 , wherein the learner is configured to generate the set of probability values for each unclassified data point using latent semantic indexing.