`(10) Patent No.:
`US 7,219,078 B2
`Lamont et al.
`(45) Date of Patent:
`May 15, 2007
`
`
`US007219078B2
`
`(54) SPATIALLY-BASED VALUATION 0F
`PROPERTY
`
`(75)
`
`Inventors: Ian G. Lamont, Coleraine (GB); David
`J McMullan Articlave (GB)
`
`(73) Assignee: Causeway Data Communications
`Limited, Coleraine (GB)
`
`( * ) Notice:
`
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(1)) by 701 days.
`
`(21) Appl. No.: 09/947,709
`
`(22)
`
`Filed:
`
`Sep. 6, 2001
`
`(65)
`
`Prior Publication Data
`
`US 2003/0046099 A1
`
`Mar. 6, 2003
`
`(5 1)
`
`Int. Cl'
`(200601)
`G06Q 40/00
`(52) US. Cl.
`............................................. 705/36; 705/1
`(58) Field Of Classification Search ------------ 705/25740,
`.
`.
`.
`705/1
`See appl1cat1on file for complete Search h1story-
`.
`References Cited
`US. PATENT DOCUMENTS
`
`(56)
`
`9/1989 Tornetta ........................ 705/1
`4,870,576 A *
`.
`11/1994 Jostetal.
`705/36
`5,361,201 A *
`
`5/1995 Hough ........
`705/10
`5,414,621 A *
`
`5,680,305 A * 10/1997 Apgar, 1v
`705/36
`1/1999 Dugan ......................... 705/36
`5,857,174 A *
`
`6,038,554 A *
`6,058,369 A *
`6,115,694 A *
`
`3/2000 Vig ............................. 705/36
`
`5/2000 Rothstein ..........
`705/36
`..... 705/36
`9/2000 Cheetham et al.
`
`
`..... 705/36
`6,141,648 A * 10/2000 Bonrssone et al.
`............ 705/36
`6,401,070 B1*
`6/2002 McManus et al.
`FOREIGN PATENT DOCUMENTS
`
`*
`
`7/2001
`
`2001188836 A
`JP
`* cited by examiner
`.
`.
`.
`.
`PrWary ExammeriFrantzy P01nv1l
`(74) Attorney, Agent, or Firm4iiford, Krass, Sprinkle,
`Anderson & Citkowski, RC.
`
`ara
`
`(57)
`
`ABSTRACT
`.
`.
`or con uc 1n
`s an me o
`are
`escr1 e
`a
`d
`d t g
`(1
`th (1
`b d f
`tu
`An pp
`spatially-based valuation of property, taking into account the
`values of neighboring properties. The method proceeds by
`analyzing a database of properties including details of at
`least the location and value of the known properties, and the
`location ofthe property to be valued. Other characteristics of
`the properties may be included in the database. Similarity
`scores for each property compared with each other property
`are calculated, and used to identify those properties with are
`comparable to the property to be valued. All the comparable
`properties within a predetermined distance from the property
`to be valued are used to calculate the unknown ro e
`p p rty
`value. The method and apparatus may also generate a report
`on the property to be valued to a user, which report may
`further include details of the comparable properties. A
`computer program for implementing the method is also
`.
`descrlbed'
`
`26 Claims, 9 Drawing Sheets
`
`1
`
`Add mmlfiqect
`
`Qv-
`
`View Rsulb [Rapm
`Bro! awn-um
`
`
`
`
`a:
`
`Minam
`
`
`
`TRULIA - EXHIBIT 1009
`
`TRULIA - EXHIBIT 1009
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 1 0f 9
`
`US 7,219,078 B2
`
`STA RT
`
`10
`
`Add Mapbase {Project
`
`.Xes OpenVectorIRasterLayers
`.p14
`
`{\b
`
`Open! ln‘port Property Dataset
`
`Renove Outliers
`
`Yes
`
`Remove Outliers &
`Update Dataset
`
`No
`
`Create Similarly
`
`
`Index
`HQ 2
`
`
`Select Area
`
`Spat'ai Con-ponent
`
`
` 28
`
`Conpute Ftoperty
`Value
`FlG 3
`
`View Results IRepod
`Error Estimation
`
`3°
`
`l%vise Criteria
`
`“5 O
`
`N?
`
`OJtput Resuts
`
`FIG 1
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 2 0f 9
`
`US 7,219,078 B2
`
`maria Values
`
` Autorrafic
`N0
`Recode
`
`
`
`38x I
`
`
`Calcdate Standardized
`Scores
`
`
`
`Create Index of
`Sin'ilarity Matrix
`
`
`Normalize Index Values
`
`46
`
`48
`
`Update Matrix
`
`'
`
`'
`
`50
`
`FIG 2
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 3 0f 9
`
`US 7,219,078 B2
`
` Select All Ursold
`Hoperties Wthin
`Radial Distance
`
`62
`
`Corrpare Unsold to
`Sold Hoperty
`
`
`
`
`
`
`
`lrdex Value
`Above Threshold
`
`Null Property
`Value
`
`64
`
`No
`
`Yes
`
`68
`No
`
`66
`
`70
`
`Move To hbxt
`thoid Hoperty
`
`Revise Property
`Value
`
`Lbdate Property Value
`
`
`7
`ast Unsold
`8 Property
`Last Sold
`
`Yes
`No
`
`Property
`
`HS 3
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 4 0f 9
`
`US 7,219,078 B2
`
` Property
`
`Attribute
`
`FIG 4
`
`
`
`
`
`_
`664999 00m
`66:19930UTEIL %
`
`244087W415111 *
`24-1504;
`5x515; 3392f
`416355
`359.85:
`244513),»
`. .w. ”um”..- .. 5...- m w. NW
`470.15:
`242817
`417724
`442.03, 29%;
`416083
`LjTBL
`nm.wu~ww~._um.§,... -WNM”... .-....,.. ..
`Ends—9'5 OUTBL:
`473,3?
`242155‘
`417702
`,
`_
`.
`
`U S. Patent
`
`May 15, 2007
`
`Sheet 5 0f 9
`
`US 7,219,078 B2
`
`ard field“
`
`CAME Bela Heleaée 1.91 ,-
`
`
`
`U.S. Patent
`
`M y 15, 2007
`
`Sheet 6 0f 9
`
`US 7,219,078 B2
`
`
`
`1 50000
`150000
`140000
`130000
`120000
`110000
`3 100000
`90000
`30000
`
`_
`
`CAMP. , Bela Release 1.01 *
`
`
`
`. Calculated Regression
`
`K COORDINATE
`
`l - :01: Surface Graph
`
`1 Cdctéated Regress 0n
`
`5 - Linear Regression
`
`'
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 7 0f 9
`
`US 7,219,078 B2
`
`
`
`1
`2
`3
`4
`5
`5
`T
`8
`9
`18
`11
`12
`13
`14
`15
`15
`17
`18
`19
`28
`21
`92
`23
`24
`25
`25
`
`1
`
`2
`
`.151
`
`.151
`.847
`.915
`.599
`.519
`.555
`.744
`.515
`.788
`.715
`.513
`.418
`.419
`.781
`.419
`.51 2
`.181
`.743
`J81
`.744
`.41?
`589
`.924
`.589
`.839
`
`.188
`.113
`.399
`.396
`.452
`.358
`.389
`.351
`.219
`.385
`.458
`.448
`.355
`.448
`.395
`.355
`.351
`.354
`.288
`.44?
`.454
`.388
`.454
`.42?
`
`3
`
`.84?
`.188
`
`5
`
`4
`
`.915
`.113
`.91 T
`
`.91}r
`.581
`.518
`.552
`.585
`.499
`.584
`.959
`.495
`.358
`.382
`.539
`.382
`.582
`.51'9
`.5?9
`5??
`194
`.378
`.552
`.144
`.552
`.533
`
`.555
`.541
`.538
`.538
`.533
`.54?
`.815
`.538
`.385
`.395
`.544
`.395
`.553
`.544
`.525
`.543
`.883
`.392
`.598
`.758
`.598
`.553
`
`.599
`.399
`.581
`.555
`
`.888
`.759
`.838
`.881
`.815
`.181
`.879
`.852
`.853
`.812
`.853
`.993
`.812
`.831
`.818
`.437
`.858
`.7134
`.554
`.734
`.559
`
`5
`
`.619
`.395
`.518
`.541
`.888
`
`.927
`.971
`.999
`.954
`.445
`.999
`.927
`.948
`.954
`.948
`.855
`.954
`.9?8
`.953
`.518
`.938
`.558
`.534
`.558
`.5?5
`
`7
`
`.555
`.452
`.552
`.538
`.759
`.927
`
`.943
`.919
`.932
`.418
`.91?
`.785
`.884
`.929
`.884
`.T87
`.929
`.938
`.92?
`.1519
`.888
`.449
`.553
`.449
`.588
`
`8
`
`.744
`.358
`.585
`.538
`.838
`.971
`.943
`
`.959
`.994
`.458
`.958
`.819
`.837
`.994
`.837
`.884
`.994
`1.888
`.994
`.511
`834
`.545
`.144
`.545
`151
`
`9
`
`.515
`.389
`.499
`.533
`.881
`.999
`.919
`.959
`
`.852
`.435
`1.888
`.931
`.844
`.952
`.944
`.849
`.952
`.959
`.952
`.518
`.942
`539
`.524
`.539
`.5283
`
`FIG 10
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 8 of 9
`
`US 7,219,078 B2
`
`FIG 11
`
`
`
`
`CAMP:
`
`Beta Release 1
`
`m
`
`
`
`U.S. Patent
`
`May 15, 2007
`
`Sheet 9 0f 9
`
`US 7,219,078 B2
`
`.31§. 93,”... w
`
`434.13 (
`V 234.85‘
`
`
`
`“0......
`FOCH
`
`.....
`
`.. . m, “m. .4
`5245.33
`
`
`
`
`
`‘sohums
`
`
`”a” .0.
`
`‘ "134%
`512:
`1012i
`:4
`
`
`
`US 7,219,078 B2
`
`1
`SPATIALLY—BASED VALUATION OF
`PROPERTY
`
`FIELD OF THE INVENTION
`
`The present invention relates to a method and an appa-
`ratus for valuation of property. In particular, but not exclu-
`sively,
`the invention relates to a computer-implemented
`method and a computer program for providing a spatially-
`based valuation of property. Certain aspects of the invention
`relate more specifically to mass appraisal of property values.
`
`BACKGROUND OF THE INVENTION
`
`is an estimate of value based on
`Property Appraisal
`opinion, of an adequately described property of a specific
`age, supported by presentation and analysis of appropriate
`information. In the case of domestic property an inspection
`is undertaken whereby a series of criteria are assessed. The
`combined output of the assessed criteria allow the valuer to
`produce an appraisal or an estimate of value for the property.
`In addition, a professional valuer may apply his or her own
`knowledge about the locality to further refine the assess-
`ment.
`
`Mass appraisal, as a method of systematic statistical
`assessment, requires the same criteria as single appraisals. It
`therefore follows the same principles as outlined above, the
`main difference between the two forms of assessment being
`multiplicity. As with single appraisals, mass appraisal mod-
`els are primarily judged by two core factors of transparency
`and predictive accuracy. There are various models or hybrid
`models and associated modelling techniques currently avail-
`able; many of these however might perform well in either
`one of the two factors of transparency and accuracy, but not
`both. This might be explained by issues such as the sample
`size, its distribution, the stratification used, the homogeneity
`of the area or other local environmental factors.
`
`It is commonly accepted that location of a property is the
`most important factor affecting its value. Significant differ-
`ences in value can occur over short distances, even within a
`single street. Property appraisers will
`infer a substantial
`amount of information about a property from its location,
`which ability is based largely on local knowledge and
`experience. In addition, location itself will exert an influence
`on nearby properties.
`Modelling of property values, therefore, should take into
`account the significant effect of location on property value.
`However, due to the difliculties of reproducing local knowl-
`edge of appraisers in a model, many valuation models do not
`directly take into account location when valuing properties,
`instead making use of ‘pseudo-location’ signifiers, such as
`local amenities, accessibility to services, and the like; how-
`ever, such signifiers do not directly reflect the influence of
`location, and so can be inaccurate. In addition, such models
`typically will incorporate many such pseudo-location signi-
`fiers in an attempt to minimise the errors inherent in such an
`approach; this therefore increases computer processing time,
`and requires detailed assessment of local areas thereby
`increasing the expense of generating the models.
`Other more complex models do attempt to incorporate
`location as a factor, all of which require an assessment of
`neighbourhoods or sub-markets. The housing market is a set
`of distinct but
`interrelated sub-markets, encompassing
`dwellings differentiated by one or several alternative dimen-
`sions. However, there is little consensus on whether sub-
`markets should be defined according to property character-
`istics, or based on the actual house price.
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`
`Sub-markets may be defined in a number of ways. In a
`spatial context,
`it
`is possible to create localised regions
`formed through the aggregation of units such as postal
`zones, enumeration districts, or ward boundaries. The use of
`‘political’ or other non-property based locational areas cre-
`ates problems related to boundary positioning; that is, such
`boundaries have not been drawn up on the basis of property
`values, and so do not truly reflect the effect of location on
`property values. Another approach is based on quantitative
`characteristics of the dwellings, such as house type, size,
`age, etc.; or house prices may be used to identify sub-
`markets. These traditional models however assume homo-
`
`geneity in distribution and thus density over a unit area.
`Further, no consideration can be given using this type of
`analytical method to trends which occur across the bound-
`aries of these areal units. Clusters of value which may be
`higher than average for the whole unit may be ‘lost’ during
`analysis. Locational analysis is normally encompassed by
`either assuming that the locational value is constant, or by
`sub-dividing units into more readily definable areas such as
`retail, business or financial districts and assuming each to be
`constant. In many cases the locational value of these districts
`may be accounted for through the valuers’ expert knowledge
`of the location.
`
`Although these models do attempt to incorporate the
`effect of location on property values,
`the sub-markets
`thereby defined are often unrepresentative of the actual
`sub-markets, while no attempt is made to treat the effects of
`location within a sub-market; that is, the models assume that
`the effect of location within a sub-market is identical on
`
`every property within the sub-market. This can lead to
`inaccurate valuations, and ignores the possibility of loca-
`tional trends within a sub-market. In order to go some way
`towards overcoming these difliculties, conventional models
`can only rely upon creating smaller and smaller sub-mar-
`kets, using more and more data, which will clearly increase
`computer processing time and use more computer memory.
`In addition, the problem of directly considering the effects of
`location on property values remains unaddressed.
`It is among the objects of embodiments of the invention
`to provide a valuation model whereby the effects of spatial
`location on property values may be directly incorporated
`into the model.
`
`It is further among the objects of certain embodiments of
`the invention to provide a computer-based modelling
`method and computer program which may generate property
`valuations at a lower computer processing burden than
`conventional valuation models, without sacrificing accuracy.
`
`SUMMARY OF THE INVENTION
`
`In certain embodiments of the invention, these objects are
`achieved, in part, by making use of statistical techniques to
`interpolate unknown property values at a particular location
`from known property values of a particular location. That is,
`X and Y co-ordinates (or any other location defining vari-
`ables) of all properties are known, together with values for
`some of the properties. Statistical techniques may then be
`used to calculate unknown property values on the basis of
`their spatial relationship to the known property values.
`Selected embodiments of the invention may incorporate
`additional factors into the model to improve accuracy of
`valuations.
`
`According to a first aspect of the present invention, there
`is provided a computer-implemented method of spatially-
`based valuation of a subject property, the method comprising
`the steps of:
`
`
`
`US 7,219,078 B2
`
`3
`selecting a subset of properties for comparison from a
`computer database containing characteristics of a plurality
`of properties of known value, the characteristics including at
`least a location and a value of each property, and the
`database further containing characteristics of a subject prop-
`erty of known location;
`generating an index of similarity matrix of the subset of
`properties;
`identifying spatially proximate comparable properties to
`the subject property, by means of a distance factor and an
`index of similarity threshold; and
`calculating the value of the subject property based on the
`values of the identified spatially proximate comparable
`properties.
`Thus, the present invention allows the calculation of a
`property value based on its spatial location with respect to
`the identified comparable properties.
`The term ‘index of similarity matrix’, as used herein,
`refers to a table or other data structure indicating the overall
`degree of similarity between each property and every other
`property for a selected dataset as a correlation coefficient for
`each pair of properties, the coefficient being derived from all
`selected attributes of the properties. Note that the usage
`herein differs from the known statistical ‘index of similarity’
`function, with the term as used herein referring to the
`scoring range from the correlation.
`The term ‘spatially proximate’ as used herein refers to
`properties which lie within a predetermined distance thresh-
`old or factor from a particular property.
`Preferably the calculation of the value of the subject
`property may be weighted with regard to the proximity of
`the comparable properties. That is, the value of properties
`which are immediately adjacent the subject property are
`likely to have a larger effect on the value of the subject
`property than properties farther away. Weighting of the
`calculated value accordingly may thus be used to account for
`location of a property in a manner which varies continuously
`throughout the area chosen, rather than treating the area as
`a single, homogenous region.
`There may of course be a plurality of subject properties
`included in the database. The step of subset selection may
`then also include the step of selecting a particular subject
`property or properties.
`The calculation of the value of the subject property may
`make use of an interpolation technique; that is, a technique
`which interpolates the value of unknown points from the
`value of known points. Most preferably,
`the calculation
`makes use of Kriging interpolation (also known simply as
`‘Kriging’); the skilled person will be aware of techniques for
`performing Kriging. In general, Kriging may be used to
`calibrate the method and/or to refine a calculated value,
`while the basic calculation of the value will be made using
`the index of similarity matrix of comparable properties, as
`described above. Alternative calculation techniques may be
`used if desired; however,
`it
`is believed at present
`that
`Kriging offers the best method for calculating the value in
`terms of computational burden and accuracy of calculated
`values. For example, spline or inverse distance weighted
`interpolation functions may be used.
`Preferably the location of the properties is recorded within
`the database as X and Y co-ordinates. This is a convenient
`
`to use, and may use for example, national grid
`format
`mapping references or latitude and longitude measurements
`as X and Y co-ordinates. Conversion of latitude and longi-
`tude measurements from sexagesimal into a more conve-
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`4
`
`nient format (for example, decimal representation) may be
`used. Other location formats may also be used, for example,
`polar co-ordinates.
`The value of the properties may be represented as abso-
`lute price, relative price, or price per square meter or square
`foot.
`
`Preferably the database includes characteristics of the
`properties in addition to location and value. Ideally, these
`characteristics will be those which are believed to have an
`
`effect on the value of a property; for example, age of
`building,
`type of building, size, number of bathrooms,
`central heating, garage, and so forth. Additional character-
`istics which do not affect property value may also be
`included, if desired.
`Selection of the subset of properties may be carried out
`automatically, for example by a computing device according
`to predetermined characteristics (such as within a predefined
`area), or manually by an operator. Where a manual selection
`is made, this may be by area (such as a town, or a district),
`or by non-spatial characteristics (for example, all detached
`houses), or by a combination of spatial and non-spatial
`characteristics. Reference herein to ‘a subset’ of properties
`may of course be taken to include selection of all properties
`on the database.
`
`The database may be local to the computer implementing
`the present method, or may be remote therefrom and
`accessed by for example a telecommunications network.
`The step of generating the index of similarity matrix
`comprises calculating a similarity score for each property
`(both of known and unknown value) when compared with
`every other property within the subset. The similarity score
`preferably is given as a correlation coefficient for the set of
`characteristics of each property; conventional statistical
`techniques may be used to calculate this correlation coeffi-
`cient. The index of similarity matrix may be calculated using
`all characteristics included within the database, or only a
`subset thereof. If a subset of characteristics is used, the
`method may further comprise the step of selecting the
`characteristics to be compared.
`Where characteristics to be compared include non-nu-
`merically coded characteristics (for example, ‘house type’
`may be coded as ‘bungalow’, ‘detached’, ‘terrace’, and so
`forth), the method may comprise the further step of auto-
`matically recoding non-numerical
`characteristics
`into
`numerical characteristics (for example, ‘house type’ may be
`recoded as 1, 2, 3, etc.). This allows such characteristics to
`be included within the index of similarity matrix. The
`numerical codes assigned may be arbitrary, provided that
`like properties will be assigned the same numerical code.
`Generation of the matrix preferably further comprises the
`step of standardising characteristics to be compared, such
`that each characteristic may be meaningfully compared with
`each other characteristic. For example, while the value of a
`property may range from, say, $10,000 to $100,000, the
`number of bedrooms may range only from 1
`to 5. Stan-
`dardisation of these characteristics allows meaningful com-
`parisons to be made. The calculated index of similarity may
`further be normalised,
`to provide a similarity score of
`between 0 and 1 for each property.
`During the identification of spatially proximate compa-
`rable properties step, the distance factor and index of simi-
`larity threshold may be manually set by an operator, or may
`be predetermined by a computer implementing the method.
`Alternatively, the method may further comprise the step of
`calculating optimal values for either or both of these values
`based on the database. For example, the optimal distance
`factor for accurate valuation of properties will vary depend-
`
`
`
`US 7,219,078 B2
`
`5
`ing on the precise contents of the database; the method may
`thus calculate this optimal factor based on the data to be
`used. The optimal distance factor may, for example, be
`generated by a semi-variogram model with the range deter-
`mining the distance of influence (spatial auto-correlation).
`The index of similarity threshold value may be generated by,
`for example, selecting values in the top percentile from the
`similarity matrix. The optimal value may be used automati-
`cally, or may be simply suggested to an operator.
`The method may further comprise the step of displaying
`the location of properties and/or other characteristics on a
`display. The display may further comprise a map of a
`relevant area; for example, a town street map may be
`overlaid with the location of each property in the database.
`The method may further comprise the step of reporting
`the calculated value of the subject property to an operator. A
`report may be given by any convenient output device; for
`example, on a display monitor, via a printer, by generated
`speech, by SMS or other messaging service transmitted over
`a telecommunications network, and so forth. The method
`may yet further comprise the step of reporting to an operator
`additional details of the calculation: for example, details of
`the identified comparable properties may be given, or an
`estimated reliability rating. This allows an operator to deter-
`mine on what basis the valuation has been made, so render-
`ing the method relatively transparent to a user.
`The method may further comprise the step of calculating
`a reliability rating for calculated value of a subject property.
`Conventional statistical techniques may be used to provide
`a reliability rating based on the database used.
`The method may yet further comprise the step of remov-
`ing properties with statistically outlying characteristics from
`the subset of properties. For example, if a database includes
`details of one particular property which sold for a significant
`amount above or below its actual value, then this outlier may
`skew the calculated values of subject properties. Removal of
`such properties improves the accuracy of the method. Out-
`liers may be identified automatically, or manually. Certain
`embodiments of the invention may allow for outliers to have
`their characteristics edited and returned to the subset of
`
`properties, rather than simply removed; for example, if a
`user identifies an outlier due to an incorrectly-entered value,
`they may edit the value to correct it rather than excluding the
`property from the database altogether.
`According to a second aspect of the present invention,
`there is provided a computer program product for spatially-
`based valuation of a subject property comprising computer
`program code recorded on a data carrier,
`the computer
`program code comprising:
`code for allowing the selection of a subset of properties
`for comparison from a computer database containing char-
`acteristics of a plurality of properties of known value, the
`characteristics including at least a location and a value of
`each property, and the database further containing charac-
`teristics of a subject property of known location;
`code for generating an index of similarity matrix of the
`subset of properties;
`code for identifying spatially proximate comparable prop-
`erties to the subject property, by means of a distance factor
`and an index of similarity threshold; and
`code for calculating the value of the subject property
`based on the values of the identified spatially proximate
`comparable properties.
`The data carrier may be in the form of magnetic tape,
`optical discs such as CD-ROMs, magnetic media such as
`floppy discs or hard discs, or permanent data carriers such as
`ROM chips and the like. The data carrier may still further
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`6
`comprise optical or electrical signal carriers such as optical
`fibres or communications cables.
`
`The computer program product may further comprise
`code for removal of outliers from the subset of properties.
`The computer program product may further comprise
`code for recording calculated property values in a database.
`According to a still further aspect of the present invention,
`there is provided a computer program for spatially-based
`valuation of a subject property, the computer program com-
`prising:
`code for allowing the selection of a subset of properties
`for comparison from a computer database containing char-
`acteristics of a plurality of properties of known value, the
`characteristics including at least a location and a value of
`each property, and the database further containing charac-
`teristics of a subject property of known location;
`code for generating an index of similarity matrix of the
`subset of properties;
`code for identifying spatially proximate comparable prop-
`erties to the subject property, by means of a distance factor
`and an index of similarity threshold; and
`code for calculating the value of the subject property
`based on the values of the identified spatially proximate
`comparable properties.
`According to a still further aspect of the present invention,
`there is provided a computer system for spatially-based
`valuation of a subject property, the system comprising:
`means for selecting a subset of properties for comparison
`from a computer database containing characteristics of a
`plurality of properties of known value, the characteristics
`including at least a location and a value of each property, and
`the database further containing characteristics of a subject
`property of known location;
`means for generating an index of similarity matrix of the
`subset of properties;
`means for identifying spatially proximate comparable
`properties to the subject property, by means of a distance
`factor and an index of similarity threshold; and
`means for calculating the value of the subject property
`based on the values of the identified spatially proximate
`comparable properties.
`The system may further comprise means for remotely
`accessing said computer database.
`The system may still further comprise means for gener-
`ating a report of the value of the subject property to a user.
`The report generation means may comprise a display means,
`such as a computer monitor or the like, printing means,
`sound generating means, or means for accessing a remote
`communications network, for example the internet, a tele-
`communications network, or the like.
`According to a yet further aspect of the present invention,
`there is provided a method of spatially-based valuation of a
`subject property, the method comprising the steps of:
`selecting a subset of properties for comparison from a
`database containing characteristics of a plurality of proper-
`ties of known value, the characteristics including at least a
`location and a value of each property, and the database
`further containing characteristics of a subject property of
`known location;
`generating an index of similarity matrix of the subset of
`properties;
`identifying spatially proximate comparable properties to
`the subject property, by means of a distance factor and an
`index of similarity threshold; and
`calculating the value of the subject property based on the
`values of the identified spatially proximate comparable
`properties.
`
`
`
`US 7,219,078 B2
`
`7
`According to a yet further aspect of the present invention,
`there is provided a method of providing a spatially-based
`property valuation to a customer, the method comprising the
`steps of:
`selecting a subset of properties for comparison from a
`computer database containing characteristics of a plurality
`of properties of known value, the characteristics including at
`least a location and a value of each property, and the
`database further containing characteristics of a subject prop-
`erty of known location;
`generating an index of similarity matrix of the subset of
`properties;
`identifying spatially proximate comparable properties to
`the subject property, by means of a distance factor and an
`index of similarity threshold;
`calculating the value of the subject property based on the
`values of the identified spatially proximate comparable
`properties;
`providing a report containing the calculated value of the
`subject property to a customer; and
`charging the customer in return for providing the report.
`The report may be provided by any convenient means, for
`example, on paper, by email or other communications net-
`work, on a computer or other display screen, verbally, via a
`personal digital assistant (PDA) or other portable electronic
`device, or the like.
`The subset of properties may be selected automatically in
`response to a query from a customer, or may be manually
`selected.
`
`The method may further comprise the steps of allowing
`the customer to select the subject property, and performing
`the remainder of the steps of the method without operator or
`customer intervention. This allows the method to be imple-
`mented for example via the intemet or the like, by which a
`customer may enter details of a subject property and receive
`by return a valuation report.
`The report may be of varying levels of detail; the method
`may further comprise the step of allowing the customer to
`select the desired level of detail, and charging the customer
`dependent on their selected level of detail.
`the
`Where the method involves customer interaction,
`for
`customer may interact
`in any convenient manner;
`example with a property valuer, via the internet, by means of
`a mobile communications device, such as a mobile tele-
`phone or personal digital assistant (PDA), or the like. The
`use of mobile interactions allows the method to be used as
`
`if a customer passes a
`an ‘impulse’ buy; for example,
`property which they like, they may immediately obtain a
`valuation to determine whether or not the property would be
`affordable.
`
`Charging of the customer may take place at the time of
`obtaining the report, or at a different time. For example,
`corporate customers may wish to pay a subscription to
`access the system, rather than pay for each use. The cus-
`tomer may be charged by debiting a credit or charge card, by
`deducting funds directly from a bank account, by deducting
`an amount of ‘electronic cash’ from a smart card or ‘elec-
`
`tronic wallet’ or similar device, by issuing an invoice, or by
`any other suitable means. Where the customer accesses the
`service by means of a telephone call or the like, the charge
`may be incorporated into the cost of the call; for example,
`the customer may access the service via a mobile telephone
`by dialling a ‘premium-rate’ telephone number.
`According to a still further aspect of the present invention,
`there is provided a method of providing a spatially-based
`valuation of a subject property, the method comprising the
`steps of:
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`8
`accessing a remote computer database containing charac-
`teristics of a plurality of properties of known value, the
`characteristics including at least a location and a value of
`each property, and the database further containing charac-
`teristics of a subject property of known location;
`selecting a subset of properties for comparison from the
`database;
`generating an index of similarity matrix of the subset of
`properties;
`identifying spatially proximate comparable properties to
`the subject property, by means of a distance factor and an
`index of similarity threshold; and
`calculating the value of the subject property based on the
`values of the identified spatially proximate comparable
`properties.
`According to a yet further aspect of the present invention,
`there is provided a method of providing a spatially-based
`valuation of a subject property, the method comprising the
`steps of:
`allowing a user to select a subset of properties for com-
`parison from a remote computer database containing char-
`acteristics of a plurality of properties of known value, the
`characteristics including at least a location and a value of
`each property, and the database further containing charac-
`teristics of a subject property of known location;
`generating an index of similarity matrix of the subset of
`properties;
`identifying spatially proximate comparable properties to
`the subject property, by means of a distanc