Ex Parte Franz et alDownload PDFBoard of Patent Appeals and InterferencesApr 15, 201110952501 (B.P.A.I. Apr. 15, 2011) Copy Citation UNITED STATES PATENT AND TRADEMARK OFFICE UNITED STATES DEPARTMENT OF COMMERCE United States Patent and Trademark Office Address: COMMISSIONER FOR PATENTS P.O. Box 1450 Alexandria, Virginia 22313-1450 www.uspto.gov APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO. 10/952,501 09/29/2004 Alexander Mark Franz 0026-0105 9453 44989 7590 04/18/2011 HARRITY & HARRITY, LLP 11350 Random Hills Road SUITE 600 FAIRFAX, VA 22030 EXAMINER YEN, SYLING ART UNIT PAPER NUMBER 2166 MAIL DATE DELIVERY MODE 04/18/2011 PAPER Please find below and/or attached an Office communication concerning this application or proceeding. The time period for reply, if any, is set in the attached communication. PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________ BEFORE THE BOARD OF PATENT APPEALS AND INTERFERENCES ____________ Ex parte ALEXANDER MARK FRANZ and MONIKA H. HENZINGER ____________ Appeal 2009-009993 Application 10/952,501 Technology Center 2100 ____________ Before JOSEPH L. DIXON, HOWARD B. BLANKENSHIP, and JOHN A. JEFFERY, Administrative Patent Judges. JEFFERY, Administrative Patent Judge. DECISION ON APPEAL Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s rejection of claims 1-29. We have jurisdiction under 35 U.S.C. § 6(b). We affirm-in-part. STATEMENT OF THE CASE Appellants’ invention monitors and improves search results using user feedback by determining a quality indication of items during a time period Appeal 2009-009993 Application 10/952,501 2 using a time-series1 model. See generally Spec. ¶¶ 0005-06, 0019. Claim 1 is reproduced below with the key disputed limitations emphasized: 1. A method comprising: providing items during a time period; and determining an indication of quality of the items provided during the time period using a time series model. The Examiner relies on the following as evidence of unpatentability: Nesamoney US 2002/0156791 A1 Oct. 24, 2002 Odom US 2006/0004732 A1 Jan. 5, 2006 (filed Aug. 2, 2005 and claiming priority to Provisional App. No. 60/592,404, filed Aug. 2, 2004, and Continuation-in-Part Application No. 10/086,026, filed Feb. 26, 2002) Steven M. Beitzel et al., Hourly Analysis of a Very Large Topically Categorized Web Query Log, SIGIR ’04: Proc. of the 27th Ann. Int’l ACM SIGIR Conf. on Res. and Dev. in Info. Retrieval 321-28 (2004) (“Beitzel”). 1 Although Appellants do not hyphenate the term “time series,” we nonetheless hyphenate it for clarity and grammatical form. Appeal 2009-009993 Application 10/952,501 3 THE REJECTIONS 1. The Examiner rejected claims 1, 4, 6-8, 14-19, 24, and 26-29 under 35 U.S.C. § 102(a) as anticipated by Beitzel. Ans. 3-7.2 2. The Examiner rejected claims 2, 3, 5, and 25 under 35 U.S.C. § 103(a) as unpatentable over Beitzel and Odom. Ans. 7-9. 3. The Examiner rejected claims 9-13 and 20-23 under 35 U.S.C. § 103(a) as unpatentable over Beitzel and Nesamoney. Ans. 9-11. THE ANTICIPATION REJECTION OVER BEITZEL Claims 1 and 6 Regarding independent claim 1, the Examiner finds that Beitzel discloses all recited limitations, including mapping Beitzel’s discussion of popularity, similarity, and frequency distributions across the hours of a day to determining a quality indicator of items during a time period using a time- series model. See Ans. 3-4, 16. Appellants argue that Beitzel does not disclose an indication of quality as recited, and that popularity and uniqueness of categorized queries are not quality indicators. App. Br. 7-10; Reply Br. 4-5. Appellants also assert that Beitzel does not mention a time-series model. App. Br. 10-14. The issue before us, then, is as follows: 2 Throughout this opinion, we refer to (1) the Appeal Brief filed September 5, 2008; (2) the Examiner’s Answer mailed November 24, 2008; and (3) the Reply Brief filed January 26, 2009. Appeal 2009-009993 Application 10/952,501 4 ISSUE Under § 102, has the Examiner erred in rejecting claim 1 by finding that Beitzel discloses determining a quality indicator of items during a time period using a time-series model? FINDINGS OF FACT (FF) 1. Appellants describe various metrics obtained that reflect the quality of search results. These metrics include the percentage of searches: (a) in which the user selected the first result in a search result list; (b) that had long clicks; and (c) that did not have another search within a short period of time. Spec. ¶ 0029. 2. Appellants state that time-series models are well known in the art and “will not be described in further detail[.]” The time-series model can reflect trends of a selected metric during the time periods in which the data is obtained and can be used to predict user behavior at a desired time. Spec. ¶¶ 0030, 0038. 3. Beitzel studies and analyzes query traffic by examining changes in popularity and uniqueness of topical queries within a general purpose commercial web search service across the hours of a day. Beitzel, Abstract; 321-327; Fig. 1. 4. Beitzel determines the relative popularity or ranking of queries within a query set by measuring the Pearson correlation of the queries’ frequency over different hours of the day. Equation 3 defines a Pearson correlation of query frequencies between hours A and B, and includes Appeal 2009-009993 Application 10/952,501 5 variables q and qi (i.e., query or queryi). Figure 2 shows four frequency distributions between different hours of the day (e.g., 12AM-1AM, 6AM- 7AM, 12PM-1PM, 6PM-7PM). Beitzel, at 323-24; Fig. 2. 5. Beitzel discusses determining popularity changes made throughout the day. Figure 8 plots a categorical (e.g., entertainment, games, health) percent of the total query volume over time (i.e., Hour 0 to 24 of the day). Beitzel, at 325; Fig. 8. 6. In order to quantify the categories’ popularity changes throughout the day, Beitzel discusses calculating a KL-divergence (Equation 4). Figure 9 plots a category percentage between 0 and 6% of the entire query stream and divergence from the likelihood of any query at each hour. Beitzel. at 325; Fig. 9. 7. Beitzel’s K-L divergence equation (Equation 4) defines a KL- divergence of query occurrence likelihood for category c and total stream at time t for query q and is reproduced below: q tcqp tqptqptcq ptq p D ),|( )|(log )|( ),|( ) | ( ( . Beitzel, at 324-25. ANALYSIS Claims 1 and 6 We begin by construing the key disputed limitation of representative claim 1 which calls for, in pertinent part, determining a quality indicator of items during a time period using a time-series model. In particular, Appellants focus on the terms “quality” and “time series model” in asserting that Beitzel does not disclose the recited determining step. However, since Appeal 2009-009993 Application 10/952,501 6 Appellants do not define these terms in the Specification (see FF 1-2), we therefore give these terms their broadest reasonable construction in light of the Specification. See In re Am. Acad. of Sci. Tech Ctr., 367 F.3d 1359, 1364 (Fed. Cir. 2004) (internal citations omitted). At the outset, the term “quality” is inherently subjective, and depends on individual subjective assessments that can vary widely regarding the nature and extent of “quality”—actual and perceived. The term “quality” is therefore quite broad, and determining this subjective characteristic is likewise quite subjective absent some objective criteria. For example, one may measure “quality” based on how many people use—or even prefer—an item (i.e., its relative popularity). Nevertheless, claim 1 recites even more broadly determining an indication of quality—not “quality” alone. Unlike a quality determination, this limitation merely requires that this determination indicates quality, which can include a variety of indicators such as the item’s usage, consumption, or popularity indicating how many people prefer a particular item. Even users’ perceptions regarding an item can “indicate” its quality— at least subjectively. Contrary to Appellants’ assertions (App. Br. 7-10), we therefore find the Examiner’s construction of “an indication of quality” broadly but reasonably includes popularity. Additionally, Appellants do not explain “a time-series model,” apart from merely indicating that this model is well known in the art. See FF 2. On appeal, Appellants provide no further explanation in this regard. See App. Br. 7-14; Reply Br. 4-6. Given Appellants’ minimal explanation, we therefore construe a “time-series model” broadly. See Am. Acad., 367 F.3d at 1364. A “model” includes “[a] mathematical or graphical representation Appeal 2009-009993 Application 10/952,501 7 of a real-world situation or object[.]”3 Thus, determining item quality using the claimed “time-series model” need be no more than determining a quality indicator during a time period using a time-series graphical or mathematical representation of real-world data. Beitzel discloses studying query logs to examine changes in popularity and uniqueness of categorical queries across the hours of a day. See FF 3. Additionally, Beitzel determines popularity or ranks queries by measuring a correlation based on a time series (e.g., range of hours) which fully meets “a time-series model” given its broadest reasonable construction. See FF 4. And as noted above, popularity or ranking queries (e.g., “items”) indicates their quality. That is, a popular item reflects that it is preferred, and therefore at least indicates that it has a higher “quality” (or at least perceived higher quality) than other, less-popular items. Additionally, the Examiner and Beitzel discuss determining the popularity changes made throughout the day. See Ans. 16; see also FF 5. For example, Beitzel’s Figure 8 determines the popularity of categories during the hours of the day or a quality indicator of items during a time period. See FF 5. Beitzel’s Figure 9 also determines a KL divergence of query occurrence within these topical categories that is also based on series of time. See FF 5-7. Based on these teachings, we therefore find that Beitzel determines a quality indicator of items provided during a time period using a time-series model as required by claim 1. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of: independent claim 1 and claim 6, not separately argued with particularity (App. Br. 7-14). 3 Microsoft® Computer Dictionary 344 (5th ed. 2002). Appeal 2009-009993 Application 10/952,501 8 Claim 4 Regarding claim 4, the Examiner finds that Beitzel discloses all recited limitations, including a query being a metric that is based on a user action. See Ans. 4, 18. Appellants argue that Beitzel’s frequency distribution (or any element in Figure 7) is not a metric for the time-series model as claimed. App. Br. 14-15; Reply Br. 6-7. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 4 by finding that Beitzel discloses selecting a metric for the time-series model that is based on a user action? ANALYSIS Based on the record before us, we find no error in the Examiner’s anticipation rejection of claim 4. As discussed above, Beitzel determines popularity (i.e., a quality indicator) which is a correlation based on a time- series model. See FF 4-7. This correlation involves variables q and qi (i.e., a query variable or metric). See id. Because a query is based on users’ searches (e.g., actions), Beitzel’s determination therefore involves selecting a metric for the model that is based on user actions. Additionally, the determined popularity trends of categories (c) are based on queries (e.g., q) or user actions. See FF 5-7. Thus, Beitzel discloses selecting another metric (e.g., categories) for modeling a time series based on a user action (e.g., searching) as required by claim 4. Appeal 2009-009993 Application 10/952,501 9 For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 4. Claim 7 Regarding claim 7, the Examiner finds that Beitzel’s Figure 8 and the disclosed predictive algorithms meet the recited determining step including predicting a metric value range for the time period using a time-series model. See Ans. 5, 19-20. Appellants reiterate that measuring popularity is not a quality measurement. App. Br. 16. Appellants also assert that Beitzel’s discussion of percentages and the Examiner’s allegedly piecemeal interpretation of Beitzel do not predict a range. App. Br. 17-18. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 7 by finding that Beitzel discloses the determining step includes predicting a metric value range? ANALYSIS Based on the record before us, we find no error in the Examiner’s anticipation rejection of claim 7. We refer to our previous discussion of claim 1. Particularly, we found that Beitzel discloses in Figures 8 and 9 determine a quality indication of items using a time series model (e.g., how popular various categorical queries are using percentages). See FF 4-7. Additionally, Figure 9 involves determining the likelihood of receiving a categorical query (e.g., predicting a metric for that time period) within a Appeal 2009-009993 Application 10/952,501 10 percentage range (e.g., 0 to 6 %). See FF 6-7. Thus, Beitzel discloses the determining step also involves predicting a metric value range, and Appellants have not persuaded us of error. We therefore sustain the rejection of claim 7. Claim 8 As to claim 8, the Examiner maps Beitzel’s Figure 9 and accompanying discussion to the determining step involves determining an actual metric value for a time period, and comparing the actual metric value to the predicted range of metric values. See Ans. 5, 21. Appellants reiterate that Beitzel does not disclose a quality indicator or predicting a metric value range. See App. Br. 18-19. Appellants also contend that the likelihood of a receiving query shown in Figure 9 is not equivalent to a predicted range. App. Br. 19. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 8 by finding that Beitzel discloses the determining step includes determining an actual metric value for the time period, and comparing the actual metric value to the predicted range of metric values? ANALYSIS Based on the record before us, we find no error in the Examiner’s anticipation rejection of claim 8. As noted previously regarding claims 1 and 7, Beitzel’s determining a quality indication of items during a time period involves predicting a metric value range. See FF 6. Additionally, Appeal 2009-009993 Application 10/952,501 11 Figure 9 uses Equation 4 to determine a KL divergence of query occurrence likelihood for a category and total stream. See FF 4-7. Since this equation (1) involves determining actual metrics (e.g., q or query and c or categories), and (2) calculating predicted ranges of metric values (e.g., summation of probabilities) that are compared (e.g., conditional probability or p(q | t) involves comparing) (see FF 7), the determining step also involves determining an actual metric and comparing the actual metric to the predicted range of metric values. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 8. Claim 14 Independent claim 14 recites two “means for” limitations with functional recitations similar to those recited in claim 1. The Examiner refers to claim 1’s discussion when rejecting claim 14. See Ans. 5, 22. Appellants argue that, unlike claim 1, claim 14 recites “means for” providing and automatically determining limitations, and the Examiner has not established a prima facie basis for denying patentability. App. Br. 20-21. Appellants also contend that Beitzel’s discussion on page 327 does not address claim 14’s features. See id. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 14 by finding that Beitzel discloses (1) means for providing items to a device, and (2) Appeal 2009-009993 Application 10/952,501 12 means for automatically determining an indication of quality of the provided items using data associated with previously-provided items? ADDITIONAL FINDINGS OF FACT 8. Appellants map the “means for providing items to a device” and “means for automatically determining an indication of quality of the provided items” limitations in claim 14 to a server (e.g., 120) or processor (e.g., 220). See App. Br. 3-4. Appellants likewise map the “means for recording user behavior,” “means for predicting user behavior,” and “means for determining an indication of quality of the items provided” in claim 28 to a server (e.g., 120) or processor (e.g., 220). See App. Br. 3-4, 6. ANALYSIS Based on the record before us, we find no error in the Examiner’s rejection of claim 14 which calls for, in pertinent part, means for providing items to a device and a means for automatically determining an indication of quality of the provided items using data associated with previously provided items. We note that neither Appellants nor the Examiner indicates whether the “means for” language found in claim 14 invokes 35 U.S.C. § 112, ¶ 6. Nevertheless, we presume that the claim invokes 35 U.S.C. § 112, ¶ 6 in view of this “means for” language. See Biomedino L.L.C. v. Waters Techs. Corp., 490 F.3d 946, 950 (Fed. Cir. 2007). We therefore construe these limitations to cover the corresponding structure described in the Specification and its equivalents. In re Donaldson Co., Inc., 16 F.3d 1189, 1195 (Fed. Cir. 1994) (en banc). Appeal 2009-009993 Application 10/952,501 13 At the outset, we refer to our discussion of claim 1 regarding how the recited functional limitations are met by Beitzel. Additionally, Appellants map both these “means for” limitations to a server (e.g., 120) or processor (e.g., 220). See FF 8. Thus, we consider these components to be the structure that performs the functions recited in claim 14. Beitzel uses a commercial web search service system to analyze various query traffic. See FF 3. Since Beitzel’s search service uses the web, it follows that a computer and its corresponding processor are used, at least in part, to provide, collect, and analyze query data. See id. That is, there are insubstantial differences between the prior art element (e.g., a computer with a processor) and the corresponding element disclosed in the Specification (e.g., a processor 220 (see FF 8)). Notably, Appellants have not shown with particularity how or why Beitzel’s web search service system that uses a computer having a processor does not perform the recited functions in substantially the same way and produces substantially the same results of providing items and automatically determining a quality indication of the items. We are therefore not persuaded of error in this equivalence, nor have Appellants demonstrated that this mapped component of Beitzel is not equivalent to the structure, material or acts described in the Appellants’ Specification. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 14. Claim 15 Independent claim 15 recites a server having memory and a processor to execute instructions that are similar in scope to steps recited in claim 1. Appeal 2009-009993 Application 10/952,501 14 The Examiner states the limitations of claim 15 are similar in scope to claim 1 and refers to claim 1’s discussion. See Ans. 5-6. Appellants argue that the Examiner ignored claim 15’s features, including a server that has a processor. App. Br. 21-22; Reply Br. 7-8. Appellants also repeat that Beitzel’s popularity and uniqueness discussion does not address claim 15’s features. App. Br. 22-23. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 15 by finding that Beitzel discloses a server with a memory and processor that executes instructions to (1) provide items to user devices, and (2) determine an indication of quality of the items using a time-series model? ANALYSIS Based on the record before us, we find no error in the Examiner’s anticipation rejection of claim 15. At the outset, we refer to our discussion of claims 1 and 14 regarding how Beitzel’s commercial web search service system that analyzes various query traffic will necessarily include a computer with a processor to perform the recited providing and determining limitations. Additionally, we find that a computer will necessarily include memory to store instructions so that the processor has the ability to execute the instructions. Lastly, claim 15 includes a preamble that recites a server. Nothing in the claim, however, precludes Beitzel’s computer from being a server in view of the term’s scope and breadth. Nevertheless, since the recited server Appeal 2009-009993 Application 10/952,501 15 is only recited in the preamble of claim 15, it does not breathe life and meaning into the claim, let alone materially affect the subsequently-recited structure or functionality. Thus, the preamble is not limiting. See American Medical Systems, Inc. v. Biolitec, Inc., 618 F.3d 1354, 1358-59 (Fed. Cir. 2010). We therefore disagree with Appellants that Beitzel does not disclose the limitations in claim 15. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 15. Claim 16 Claim 16 recites a computer-readable memory device containing instructions to perform a method for performing the steps similar to those in claim 1. The Examiner indicates that the limitations of claim 16 are similar in scope to claim 1, and refers to claim 1’s discussion. See Ans. 6, 26. Appellants repeat the arguments presented for claim 1. Compare App. Br. 23-30 with App. Br. 7-14. The issues before us, then, are similar to those in connection with claim 1. We are not persuaded for the reasons previously discussed in connection with claim 1. Regarding the preamble’s reciting a “memory device,” we find, similar to our findings regarding claim 15, that this preamble limitation does not breathe life and meaning into the claim, let alone materially affect the subsequently-recited structure or functionality. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 16. Appeal 2009-009993 Application 10/952,501 16 Claims 17-19 and 24 Representative independent claim 17 recites a method that (1) records and predicts user behavior relating to items, and (2) determines a quality indicator of the items based on the user behavior and the predicted behavior. The Examiner relies on the discussion regarding claims 1, 4, 7, and 8 to reject this claim. See Ans. 6. Appellants argue that Beitzel’s Figure 8 and the popularity discussion do not predict user behavior for time period. App. Br. 30-32. The issues before us, then, are similar to those in connection with claims 1 and 7. ANALYSIS Based on the record before us, we find no error in the Examiner’s rejection of claim 17. At the outset, we refer to our previous discussion of claims 1 and 7 regarding how Beitzel discloses predicting user behavior (e.g., likelihood of receiving a categorical query). We therefore are not persuaded that the Beitzel’s discussion of popularity and calculating a KL-divergence that involves determining the likelihood of receiving a query in a particular category (see FF 4-7) does not involve predicting user behavior for a time period as recited in claim 17. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claim 17, and claims 18, 19 and 24 which are not argued with particularity (see App. Br. 30-32). Appeal 2009-009993 Application 10/952,501 17 Claims 26-29 Representative independent claim 26 recites a system that includes logic to (1) provide items over time; (2) record user behavior related to the items; (3) predict the user behavior for the time period; and (4) determine a quality indicator of the items based on user behavior and predicted user behavior. The Examiner relies on the discussion regarding claims 1 and 17. See Ans. 6. Appellants’ arguments are similar to those for claims 1 and 17, namely that Beitzel does not disclose logic to provide items over time or logic to predict user behavior. See App. Br. 32-34. The issue before us, then, is as follows: ISSUE Under § 102, has the Examiner erred in rejecting claim 26 by finding that Beitzel discloses a system having logic to (1) provide items over a time period, and (2) predict user behavior over the time period? ANALYSIS Based on the record before us, we find no error in the Examiner’s rejection of claim 26. We refer to our previous discussion of claims 1 and 17 regarding how Beitzel discloses providing items over time and predicts user behavior (e.g., likelihood of receiving a categorical query) for the time period. See FF 3-7. We therefore are not persuaded that the Beitzel does not disclose logic to perform these functions as recited in claim 26. The scope of claims 27 and 28 is similar to those of claim 26. Claim 27 recites a memory having instructions for performing steps substantially similar to the logic in claim 26. Claim 28 recites “means for” limitations Appeal 2009-009993 Application 10/952,501 18 that perform the recording, predicting, and determining logic steps in claim 26. The Examiner relies on claims 1 and 17’s discussion to reject these claims. See Ans. 6-7. Since Appellants’ arguments for claims 27-29 are substantially similar to those for claim 26 (compare App. Br. 32-34 with App. Br. 34-39), we are not persuaded for the reasons previously discussed. Additionally, while claim 28 invokes 35 U.S.C. § 112, ¶ 6, Appellants have not demonstrated that Beitzel’s components (e.g., processor) are not equivalent to the structure, material or acts described in the Appellants’ Specification. See App. Br. 6. For the foregoing reasons, Appellants have not persuaded us of error in the anticipation rejection of claims 26-29. THE OBVIOUSNESS REJECTION OVER BEITZEL AND ODOM Claim 2 Regarding claim 2, the Examiner finds that Beitzel discloses all recited limitations, except that the items include search results. Ans. 7-8. The Examiner relies on Odom to cure this deficiency in concluding that the claim would have been obvious. Ans. 8-9. Appellants argue that Odom does not cure the purported deficiencies of Beitzel in claim 1. App. Br. 40. Additionally, Appellants assert that Beitzel is not concerned with search results, but rather queries. App. Br. 41. Appellants further contend that the Examiner concludes with no reasoning that Odom and Beitzel are combinable to establish a case of obviousness. App. Br. 41-42. The issues before us, then, are as follows: Appeal 2009-009993 Application 10/952,501 19 ISSUES (1) Under § 103, has the Examiner erred in rejecting claim 2 by finding that Beitzel and Odom collectively would have taught or suggested that the recited items include search results? (2) Is the Examiner’s reason to combine the teachings of these references supported by articulated reasoning with some rational underpinning to justify the Examiner’s obviousness conclusion? ADDITIONAL FINDINGS OF FACT 9. Odom’s search engine methods generate relevant search results and associate relevant advertisements with the search results. Odom’s artificial intelligence (AI) system evaluates query entries made by the user and re-evaluates as the searcher picks new topics or adds words to the search. Odom, ¶¶ 0019, 0093. ANALYSIS Based on the record before us, we find no error in the Examiner’s obviousness rejection of claim 2, which recites, in pertinent part, that the recited items include search results. As stated above regarding claim 1, Beitzel discloses all limitations, and thus we need not address whether Odom cures any alleged deficiency (see App. Br. 40). Nevertheless, attacking Beitzel individually (see App. Br. 41) does not show nonobviousness where rejections are based on the collective teachings of Beitzel and Odom. See In re Merck & Co., Inc., 800 F.2d 1091, 1097 (Fed. Cir. 1986). We are therefore not persuaded by Appellants’ argument that Appeal 2009-009993 Application 10/952,501 20 Beitzel fails to disclose the items are search results (see App. Br. 41) when the Examiner has relied upon the collective teachings of Beitzel and Odom to teach this limitation. See Ans. 7-8. In addition, the Examiner cites Odom to teach monitoring search results. See Ans. 8. The Examiner provides a reason to combine Odom with Beitzel to teach the limitations of claim 2, namely, allowing for a more efficient and flexible search engine that enabling advertisers to pinpoint appropriate advertising based on the relevancy of the search results and advertisements. See Ans. 8-9; see also FF 9. Moreover, combining Odom with Beitzel would have predictably yielded a method that not only monitors or determines a quality indicator based on the query (see FF 3-7), but also the search results and advertisements (see FF 9) in order to advertise to users effectively. See KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398, 417 (2007). Thus, we disagree with Appellants that the Examiner only provides a mere conclusory statement to support the conclusion of obviousness. App. Br. 41. For the foregoing reasons, Appellants have not persuaded us of error in the Examiner’s rejecting claim 2. Claim 3 Regarding claim 3, the Examiner finds that Beitzel discloses all recited limitations, except that the items include advertisements. Ans. 7-8. The Examiner relies on Odom to cure this deficiency. Ans. 8-9. Appellants make similar arguments to claim 2, namely that (1) Odom does not cure the purported deficiencies of Beitzel related to claim 1, and (2) the Examiner only made a conclusory statement to support the conclusion of obviousness. App. Br. 42-44. Additionally, Appellants assert that Beitzel is Appeal 2009-009993 Application 10/952,501 21 not concerned with advertisements, and combining Odom’s teaching related to advertising with Beitzel amounts to impermissible hindsight. App. Br. 42-43. The issues before us, similar to claim 2, are as follows: ISSUES (1) Under § 103, has the Examiner erred in rejecting claim 3 by finding that Beitzel and Odom collectively would have taught or suggested the recited items include advertisements? (2) Is the Examiner’s reason to combine the teachings of these references supported by articulated reasoning with some rational underpinning to justify the Examiner’s obviousness conclusion? ANALYSIS Based on the record before us, we find no error in the Examiner’s obviousness rejection of claim 3, which recites, in pertinent part, that the recited items include advertisements. We are not persuaded by Appellants’ arguments for the reasons set forth for claim 2. Also, we note that Odom teaches evaluating search results, which include advertisements, and thus also effectively evaluates advertisements. See FF 9. Moreover, as discussed above regarding claim 2, we are not persuaded that the Examiner has relied upon impermissible hindsight in formulating the obviousness rejection. See App. Br. 42-43. For the foregoing reasons, Appellants have not persuaded us of error in the rejection of claim 3. Appeal 2009-009993 Application 10/952,501 22 Claims 5 and 25 Representative claim 5 depends from claim 4. The Examiner finds that Beitzel discloses many of the recited limitations in claim 5, including the metric comprising one or more of a percentage of times that a high ranking search is selected, an average first click position, a percentage of searches that had long clicks, a percentage of searches that did not have another search within a first period of time, or a percentage of searches that did not have a reformulated search within a second period of time. Ans. 7-8. The Examiner relies on Odom to teach the items comprise search results. Ans. 8-9. Appellants argue that Odom does not cure the purported deficiencies of Beitzel related to claim 4. App. Br. 44. Additionally, Appellants assert that Beitzel and Odom do not teach the recited metric. App. Br. 45-46. Specifically, Appellants argue that Beitzel’s discussion of the reformulated queries being 40-52% of queries in a log does not teach a percentage of searches that did not have reformulated search within a second time period. App. Br. 45-46. Appellants also repeat that the Examiner concludes without some rational underpinning that Odom with Beitzel are combinable. App. Br. 46. The issues before us, then, are as follows: ISSUES (1) Under § 103, has the Examiner erred in rejecting claim 5 by finding that Beitzel and Odom collectively would have taught or suggested a metric that comprises percentage of searches that did not have a reformulated search within a second time period? Appeal 2009-009993 Application 10/952,501 23 (2) Is the Examiner’s reason to combine the teachings of these references supported by articulated reasoning with some rational underpinning to justify the Examiner’s obviousness conclusion? ADDITIONAL FINDINGS OF FACT 10. Beitzel describes various studies that focus on user behavior at a query session level with varying results, with some estimating reformulated queries constitute 40-52% of queries in a log. Beitzel, at 322. ANALYSIS Based on the record before us, we find no error in the Examiner’s obviousness rejection of claim 5, which recites, in pertinent part, the items are search results and a metric that comprises percentage of searches that did not have reformulated search within a second time period. As stated above regarding claim 4, Beitzel discloses all recited limitations, and thus we need not address whether Odom cures any alleged deficiency (see App. Br. 44). Also, when addressing claim 2, we noted how Odom teaches items that comprise search results, and how the Examiner articulated reasoning with some rational underpinning to justify the Examiner’s combining Beitzel and Odom in concluding the claim would have been obvious. Regarding the limitation to a metric that comprises percentage of searches that did not have reformulated search within a second time period, Beitzel teaches studying user behavior and, in particular, examining the percentage of reformulated queries in a query log. FF 10. As the Examiner states (Ans. 7-8), Beitzel at least suggests studying metrics in addition to search queries, including a reformulated query percentage. See id. Appeal 2009-009993 Application 10/952,501 24 Furthermore, an ordinarily skilled artisan would have known how to measure a percentage of searches that did not reformulate by subtracting the reformulated percentage from 100. Lastly, Beitzel’s reformulated percentage must inherently involve some period (e.g., a second time period) separate from the time period used to provide the items. We therefore disagree with Appellants (App. Br. 45) that Beitzel does not teach or suggest the limitations in claim 5. For the foregoing reasons, Appellants have not persuaded us of error in the rejection of claim 5 and claim 25, which presents similar arguments to claim 5 (compare App. Br. 47-49 with App. Br. 44-46). THE OBVIOUSNESS REJECTION OVER BEITZEL AND NESAMONEY Claim 9 Claim 9 depends from claim 8. The Examiner finds that Beitzel discloses all recited limitations in claim 9, except for taking remedial measures when the actual metric for the time period falls outside the predicted range of metric values. Ans. 9-10. The Examiner relies on Nesamoney to cure this deficiency. Ans. 10-11. Appellants argue that Nesamoney does not cure the purported deficiencies of Beitzel in claim 8. App. Br. 49. Additionally, Appellants assert that Beitzel does not teach taking remedial measures as recited in claim 9, and that Nesamoney does predict a range of metric values as recited in claim 7. App. Br. 50-51. Appellants further contend that the Examiner has not provided a reason with some rational underpinning to find the combination of Odom and Beitzel supports a conclusion of obviousness. App. Br. 52. Appeal 2009-009993 Application 10/952,501 25 The issues before us, then, are as follows: ISSUES (1) Under § 103, has the Examiner erred in rejecting claim 9 by finding that Beitzel and Nesamoney collectively would have taught or suggested taking remedial measures when the actual metric value for a time period falls outside the predicted range of metric values? (2) Is the Examiner’s reason to combine the teachings of these references supported by articulated reasoning with some rational underpinning to justify the Examiner’s obviousness conclusion? ADDITIONAL FINDINGS OF FACT 11. Nesamoney teaches a technique for organizing data that includes a “push” approach. The approach can include a user specifying thresholds or conditions under which a user should be notified or alerted of a change in the data of interest, including if a metric changes by a certain amount or a certain percentage. Nesamoney, Abstract; ¶ 0047. ANALYSIS Based on the record before us, we find no error in the Examiner’s obviousness rejection of claim 9, which recites, in pertinent part, taking remedial measures when the actual metric value for a time period falls outside the predicted range of metric values. As stated above regarding claims 7 and 8, Beitzel discloses all recited limitations, and thus we need not address whether Nesamoney cures any alleged deficiency (see App. Br. 49). Moreover, attacking Beitzel or Nesamoney individually (see App. Br. 51-52) Appeal 2009-009993 Application 10/952,501 26 does not show nonobviousness where the rejection is based on the collective teachings of Beitzel and Nesamoney. See Merck, 800 F.2d at 1097. We are therefore not persuaded by Appellants’ argument that Beitzel fails to teach taking remedial measure steps (see App. Br. 51) or that Nesamoney fails to teach predicting a range of metric values, for the Examiner relies upon the collective teachings of Beitzel and Nesamoney to teach these limitations. See Ans. 9-11. In addition, the Examiner cites (1) Beitzel to teach determining a quality indicator by ascertaining actual and predicted metric values and comparing the values, as discussed above in connection with claim 8, and (2) Nesamoney to teach taking remedial steps. See Ans. 9-10. Moreover, the Examiner provides a reason to combine Nesamoney with Beitzel (see Ans. 10-11), namely to allow the system to access information of particular interest to users in large databases more efficiently. See id; see also FF 11. Also, combining Odom with Beitzel would have predictably yielded a method that takes remedial measures by notifying a user of significant changes in actual metric values outside a predicted range or delta (e.g., search percentage within a certain category falls outside predicted percentage range (see FF 5-7, 11)). See KSR, 550 U.S. at 417. Thus, we disagree with Appellants that the Examiner has presented a conclusory statement in concluding the claim would have been obvious over these collective teachings. App. Br. 52. Lastly, we disagree with Appellants that the Examiner has broken the claim language into illogical portions. See App. Br. 50. While the Examiner separates claim language by mapping different parts of the claim to Beitzel, the Examiner has nonetheless considered the claim in its entirety. That is, Appeal 2009-009993 Application 10/952,501 27 for clarity, the Examiner’s rejection makes side-by-side comparisons with various recited limitations and the relevant passages from the cited references teaching or suggesting these limitations, and provides examples at appropriate portions of the claim. See Ans. 9-10. For the foregoing reasons, Appellants have not persuaded us of error in the rejection of claim 9. Claim 10 Claim 10 depends from claim 1. The Examiner finds that Beitzel teaches all recited limitations in claim 10. See Ans. 9-10. Appellants refer to the arguments made for claim 1, and argue that Nesamoney does not cure the alleged deficiencies. App. Br. 52-53. We are not persuaded of error for the reasons discussed above in connection with claim 1. Claims 11 and 12 Representative claim 11 depends from claim 1. The Examiner finds that Beitzel teaches all recited limitations in claim 11, except for taking remedial measures. See Ans. 9-10. The Examiner relies on Nesamoney to cure the deficiency. See Ans. 10-11. Appellants refer to the arguments made for claim 1, and contend that Nesamoney does not cure the purported deficiencies. App. Br. 53. We are not persuaded by these arguments, however, for the reasons discussed previously in connection with claims 1 and 9. We therefore sustain the rejection of claim 11, and claim 12 not argued with particularity (see App. Br. 53). Appeal 2009-009993 Application 10/952,501 28 Claims 20-22 Representative claim 20 depends from claim 19, and recites (1) comparing a value representing the recorded user behavior to the determined range of predicted values, and (2) determining that the quality indication is below an expectation when the value is outside the range of predicted values. The Examiner finds the scope of claim 20 is similar to claims 8, 9, and 11-13 and refers to these rejections. See Ans. 11. Appellants present arguments similar to those in connection with claims 8, 9, and 11-13 (see App. Br. 53-56), including: (1) Nesamoney does not cure the purported deficiencies of Beitzel; (2) Beitzel and Nesamoney do not teach (a) determining a quality indicator of items involves comparing a value representing a recorded behavior to a determined range of predicted values, or (b) determining when a quality indicator falls outside a predicted range; and (3) the Examiner has not articulated a reason with some rational underpinning to combine Odom with Nesamoney and support a conclusion of obviousness. See id. As the scope of claim 20 has similarities to claims 8, 9, and 11, the issues before us are similar to those in connection with claims 8, 9, and 11. ANALYSIS Based on the record before us, we find no error in the Examiner’s obviousness rejection of claim 20, which recites, in pertinent part, comparing a value representing the recorded user behavior to the determined range of predicted values and determining the quality indication is below an expectation when the value is outside the range of predicted values. As stated above regarding claim 8, Beitzel determines actual recorded behaviors Appeal 2009-009993 Application 10/952,501 29 (e.g., a query in certain category), and compares them to predicted range of values (e.g., likely queries in different categories. See FF 6-7. We thus need not address whether Nesamoney cures this alleged deficiency. Moreover, attacking Beitzel individually (see App. Br. 54-55) does not show nonobviousness where rejections are based on the collective teachings of Beitzel and Nesamoney. See Merck, 800 F.2d at 1097. We are therefore not persuaded by Appellants’ argument that Beitzel fails to teach taking remedial measure steps when the Examiner has relied upon the collective teachings of Beitzel and Nesamoney to teach this limitation. See Ans. 9-11. Lastly, we refer to our discussion of claims 9-11 relating to Nesamoney and its teachings in finding that the Examiner’s position is based on at least some rational underpinning for combining Nesamoney with Beitzel to predictably yield the invention of claim 20. For the foregoing reasons, Appellants have not persuaded us of error in the rejection of claim 20, and claims 21 and 22 not argued with particularity (see App. Br. 53-56). Claims 13 and 23 Claim 13 depends from claim 11, and claim 23 depends from claim 21. Although claim 13 has not been separately addressed by Appellants (see App. Br 53), claim 13 recites the remedial measure comprise “automatically removing a modification associated with the providing” and is similar in scope to claim 23. The Examiner states that Beitzel discloses the removing limitation, and provides an example of returning to a predefined condition. See Ans. 20. However, the Examiner also states that Nesamoney does not disclose removing a modification associated with the provided items but Appeal 2009-009993 Application 10/952,501 30 finds that “this is simply [a] return to the predefined condition when a notification is sent.” Ans. 11. For claim 23, the Examiner relies on the previous discussion of claim 13. See id. When arguing claim 23, Appellants refer to the arguments made for claim 21. See App. Br. 56. Additionally, Appellants contend that Beitzel and Nesamoney collectively do not teach nor suggest taking remedial measures that automatically remove the system modifications. See App. Br. 56-57. The issue before us, then, is as follows: ISSUE Under § 103, has the Examiner erred in rejecting claim 13 finding that Beitzel and Nesamoney collectively would have taught or suggested the remedial measures include automatically removing a modification associated with providing the items during a time period? ANALYSIS Based on the record before us, we find error in the Examiner’s obviousness rejection of claim 13, which recites, in pertinent part, automatically removing a modification associated with providing the items during a time period. In rejecting claims under 35 U.S.C. § 103, it is incumbent upon the Examiner to establish a factual basis to support the legal conclusion of obviousness. See In re Fine, 837 F.2d 1071, 1073-74 (Fed. Cir. 1988). We find that the Examiner has not met the requisite burden to establish a factual basis to support a conclusion of obviousness for claim 13. Appeal 2009-009993 Application 10/952,501 31 When discussing Beitzel, the Examiner maps the “automatically removing a modification” step to providing an example of “return[ing] to the predefined condition.” See Ans. 10. Nonetheless, there is no explanation what the “predefined condition” is, let alone where in Beitzel (e.g., pages or paragraphs) returning to such a “predefined condition” (even assuming, without deciding, that such a condition exists in Beitzel) is taught. See Ans. 9-10. Nesamoney does discuss a threshold or condition that is used to notify a user of changes which would, at least broadly, provide a “remedial measure.” See FF 10; see also Ans. 10. Perhaps this is the “predefined condition” the Examiner is referring to when discussing Beitzel. However, the Examiner also admits that “Nesamoney fails to indicate removing a modification associated with the providing . . . .” Ans. 11. The Examiner then finds this limitation is met by stating “this is simply [a] return to the predefined condition when a notification is sent.” See Ans. 11. Again, the Examiner has provided inadequate evidence in Nesamoney to support this conclusion of obviousness, and to add a “removing a modification” step to the combined Beitzel/Nesamoney method that takes remedial measures when a quality indicator is below an expectation. Thus, Nesamoney does not reasonably teach nor suggest that removing a modification associated with the providing step should be added to the combined Beitzel/Nesamoney method of taking remedial measures. Moreover, even considering the ordinarily skilled artisan’s background knowledge, we do not find the Examiner’s explanation provides a sufficient reason to combine such a removal step with Beitzel/Nesamoney’s method of taking remedial measures. See KSR, 550 U.S. at 418. The Examiner alleges that such a combination would facilitate Appeal 2009-009993 Application 10/952,501 32 access to information in large databases and particularly items of interest. See Ans. 11. While an ordinarily skilled artisan may recognize that removing items from a database makes a database smaller, and smaller databases are easier to search, this recognition does not teach or suggest that removing a modification associated with the provided items will reduce the database’s size commensurate with the Examiner’s apparent position. We therefore find that the Examiner has not met the requisite burden of establishing obviousness for claim 13. Claim 23 has a similar “automatically removing the modification” step to claim 13, namely “wherein the remedial measures comprise automatically removing the modification to the system.” For the above reasons, we find that the Examiner has likewise failed to establish a factual basis to support the obviousness rejection of claim 23. For the foregoing reasons, Appellants have persuaded us of error in the rejection of claims 13 and 23. CONCLUSION The Examiner did not err in rejecting (1) claims 1, 4, 6-8, 14-19, 24, and 26-29 under § 102, and (2) claims 2, 3, 5, 9-12, 20-22, and 25 under § 103. The Examiner, however, erred in rejecting claims 13 and 23 under § 103. ORDER The Examiner’s decision rejecting claims 1-29 is affirmed-in-part. Appeal 2009-009993 Application 10/952,501 33 No time period for taking any subsequent action in connection with this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv). AFFIRMED-IN-PART pgc Appeal 2009-009993 Application 10/952,501 34 EVIDENCE APPENDIX Microsoft® Computer Dictionary 344 (5th ed. 2002). Notice of References Cited Application/Control No. 10/952,501 Applicant(s)/Patent Under Reexamination Examiner Syling Yen Art Unit 2100 Page 1 of 1 U.S. PATENT DOCUMENTS * Document Number Country Code-Number-Kind Code Date MM-YYYY Name Classification A US- B US- C US- D US- E US- F US- G US- H US- I US- J US- K US- L US- M US- FOREIGN PATENT DOCUMENTS * Document Number Country Code-Number-Kind Code Date MM-YYYY Country Name Classification N O P Q R S T NON-PATENT DOCUMENTS * Include as applicable: Author, Title Date, Publisher, Edition or Volume, Pertinent Pages) X U Microsoft® Computer Dictionary 344 (5th ed. 2002). V W X *A copy of this reference is not being furnished with this Office action. (See MPEP § 707.05(a).) Dates in MM-YYYY format are publication dates. Classifications may be US or foreign. U.S. Patent and Trademark Office PTO-892 (Rev. 01-2001) Notice of References Cited Part of Paper No. Delete Last PagelAdd A PageAuto-Fill Copy with citationCopy as parenthetical citation