A vocabulary recommendation method for spatiotemporal data discovery based on Bayesian network and ontologies

Figures & data

Figure 1. General workflow of the research method.

Table 1. The 20 most frequent keywords in user search history.

Keyword	Frequency	Keyword	Frequency
SEA/OCEAN	42451	RAW	5437
DAY/DAILY	27648	MONTHLY	5424
TEMPERATURE	19583	SST	5299
GLOBAL	17057	DATASET	4938
SURFACE	15994	CYCLE	4834
HISTORICAL	14000	RESOLUTION	4595
NETCDF	12802	CIRCULATION	4325
GHRSST	6958	PREVIEW	4152
WIND	6061	TOPOGRAPHY	3656
NEAR_REAL_TIME	5569	HIGH	3347

Figure 2. The general structure of the ontologies in this project.

Figure 3. Bayesian network for data discovery.

Table 2. Comparing Bayesian network model to native MUDROD recommendations.

Search Keyword	Result (Search Result/Model Inference)	Precision	Recall
WIND	Native Result: OCEAN, SURFACE, SPATIAL RESOLUTION, ICE, QUICKSCAT, DAY, NETCDF, SEAWINDS, DATASET, GLOBAL	70%	70%
	Bayesian Model Inference: AVHRR-3, TEMPERATURE, DAY, GLOBAL, SURFACE, OCEAN, NETCDF, SPATIAL RESOLUTION, SEAWINDS, GHRSST
NETCDF	Native Result: OCEAN, GRACE, SURFACE, GLOBAL, DAY, GHRSST, TEMPRATURE, JASON-1, ASCII, PRESSURE	70%	70%
	Bayesian Model Inference: OCEAN, SURFACE, WIND, GHRSST, TEMPERATURE, AVHRR-3, GLOBAL, JASON-1, HISTORICAL, DAY
OCEAN	Native Result: GLOBAL, WIND, DAY, SUFFACE, AVHRR-3, GHRSST, DATASET, MONTHLY, NETCDF, TEMPERATURE	90%	90%
	Bayesian Model Inference: GHRSST, AVHRR-3, TEMPERATURE, NETCDF, DATASET, HISTORICAL, DAY, GLOBAL, SURFACE, WIND
GRACE	Native Result: WATER, OCEAN, GLOBAL, GRAVITY, MONTHLY, DATASET, TEMPERATURE, SPATIAL RESOLUTION, SURFACE, NETCDF	70%	70%
	Bayesian Model Inference: OCEAN, SURFACE, WIND, GHRSST, AVHRR-3, TEMPERATURE, GLOBAL, NETCDF, MONTHLY, SPATIAL RESOLUTION
RADAR	Native Result: NETCDF, OCEAN, JASON-1, GLOBAL, SPATIAL RESOLUTION, SURFACE, DAY, SEAWINDS, RAPIDSCAT, SOURTHERN ASIA	80%	80%
	Bayesian Model Inference: SEAWINDS, WINDS, OCEAN, SURFACE, DAY, NETCDF, GLOBAL, TEMPERATURE, SPATIAL RESOLUTION, JASON-1

Table 3. Comparison result for query 1.

Query 1: NSCAT_IFREMER_L3_OW_MEAN_WIND_FIELDS Seed Keywords: OCEAN WIND VECTOR DATASET NETCDF Inferred Keywords: SURFACE, SPATIAL RESOLUTION, TEMPERATURE, GLOBAL, GHRSST, DAY/DAILY, HISTORICAL, AVHRR-3, MONTHLY, RAW
Evaluation	Dataset Names/Keywords	Precision
EXCELLENT	NSCAT_LEVEL_3_V2 NSCAT_LEVEL_3_BROWSE_IMAGES	86.7%
	OCEAN, WIND, VECTOR, DATASET, SURFACE, SPATIAL RESOLUTION, HDF, RAW, TEMPERATURE, GLOBAL, HISTORICAL, MONTHLY, DAY, AVHRR-3, ADEOS II
GOOD	ERS-1_IFREMER_L3_OW_MEAN_WIND_FIELDS SEAWINDS_LEVEL_3_V2	66.7%
	OCEAN, WIND, VECTOR, DATASET, SURFACE, NETCDF, SPATIAL RESOLUTION, MONTHLY, GLOBAL, DAILY, ERS-1, SEAWINDS, ICE, ADEOS II, 12 HOURS
OK	REMSS-L2P_GRIDDED_25-AMSRE QSCAT_L3_SFC_WIND_SPEED_1DEG_1MO	66.7%
	OCEAN, SURFACE, TEMPERATURE, NETCDF, DAY, GLOBAL, MONTHLY, WIND, VECTOR, HOUR, GHRSST, RESOLUTION, 12 HOURS, SEAWINDS, QUICKSAT
BAD	EUR-L2P-AVHRR17_G SEAWINDS_BYU_L3_OW_SIGMA0_ENHANCED	40%
	GHRSST, OCEAN, SURFACE, GLOBAL, NOAA-17, SST, DATASET, RESOLUTION, AVHRR-3, 12 HOURS, SEAWINDS, ADEOS II, ICE, HOUR, RADAR

Table 4. Comparison result for query 2.

Query 2: NSCAT_LEVEL_1.7_V2 Seed Keywords: RADAR, ADEOS-2, ICE, 12 HOURS, DATASET Inferred Keywords: OCEAN, SURFACE, WIND, SPATIAL RESOLUTION, TEMPERATURE, GLOBAL, NETCDF, HISTORICAL, SEAWINDS, DAY/DAILY
Evaluation	Dataset Names/Keywords	Precision
GOOD	WENTZ_SASS_SIGMA0_L2 QSCAT_LEVEL_2A_V2	73.3%
	12 HOURS, OCEAN, DATASET, RADAR, WIND, SEAWINDS, SURFACE, DAY, ICE, VECTOR, HISTORICAL, SPATIAL RESOLUTION, HOUR, QUICKSAT, RAW
OK	RSCAT_L1B_V1.1 RSCAT_L1B_V1.2	53.3%
	OCEAN, RADAR, SEAWINDS, 12 HOURS, DATASET, VECTOR, WIND, DAY, ANALYSIS, ORBIT, RAW, HDF, QUICKSAT, HOUR, RESOLUTION
BAD	JASON-1_L2_OST_GPS_E JASON-1_L2_OST_GPR_E_GEODETIC	46.7%
	OCEAN, RADAR, DAY, NETCDF, SURFACE, DATASET, RESOLUTION, WAVE, WIND, ORBIT, WATER, RAW, VECTOR, JASON-1, TOPOGRAPHY

Table 5. Comparison result for query 3.

Query 3: AVHRR_PATHFINDER_L3_SST_DAILY_DAYTIME_V51 Seed Keywords: PATHFINDER, AVHRR-3, SURFACE, TEMPERATURE, DAY/DAILY Inferred Keywords: OCEAN, SST, GLOBAL, RAW, NETCDF, HISTORICAL, GHRSST, MONTHLY, WIND, NOAA-18
Evaluation	Dataset Names/Keywords	Precision
EXCELLENT	AVHRR_PATHFINDER_L3_SST_DAILY_NIGHTTIME_V5 AVHRR_PATHFINDER_L3_SST_7DAY_NIGHTTIME_V51	80%
	PATHFINDER, OCEAN, SURFACE, TEMPERATURE, SST, AVHRR-3, DAY, GLOBAL, MONTHLY, HISTORICAL, RAW, NOAA-17, HDF, NOAA-18, DATASET
GOOD	AVHRR_PATHFINDER_L3_SST_ANNUAL_NIGHTTIME_V51 AVHRR_PATHFINDER_L3_SST_ANNUAL_DAYTIME_V51	66.7%
	PATHFINDER, OCEAN, SURFACE, TEMPERATURE, SST, AVHRR-3, ANNUAL, NOAA-18, NOAA-17, NOAA-16, MONTHLY, DATASET, HDF, DAY, GLOBAL
BAD	TELLUS_DOT_2008_DATA SEAWINDS_BYU_L3_OW_SIGMA0_ARCTIC_POLAR-STEREOGRAPHIC_BROWSE_IMAGES_LTOD	20%
	OCEAN, SURFACE, ASCII, DATASET, DAY, 12 HOURS, SEAWINDS, ADEOS-II, RESOLUTION, QUICKSAT, ICE, HOUR, RADAR, CIRCULATION, TOPOGRAPHY

Data availability statement

The data referred to in this paper is not publicly available at the current time.