Comparator and product selection method

Abstract

When choosing a product to buy, you can simply compare two products and select the one you prefer, allowing you to choose the product that best suits your purchase purpose. [Solution] An "n-th selection cycle" is performed, in which the user selects the first, second, ... nth product with the highest preference from the product information of multiple purchase candidates displayed on the screen. The selected few provisional products (first, second, ... nth) are then compared to make a final selection. In the nth selection cycle, two different products are displayed from the candidate list, and the user is asked to select one, choosing the product with the highest preference. An n+1 selection cycle is then performed from the group of products that were not preferred when directly compared with the product that was most preferred in the nth selection cycle.

Description

Technical Field

[0001] It relates to a technology for comparing and selecting items to be purchased on a computer screen.

Background Art

[0002] (For inquiries regarding this case, please contact the inventor, Akira Hirano, at 029 - 845 - 4668)

[0003] Since the present invention relates to technologies on a computer screen or a web browser, the "product" refers not to the physical object but to product information on the screen (such as photos, strings including product names, identifiers, cards, etc.).

[0004] The user's preference for a product is called "preference", and "when comparing Product A1 and Product A2 and the preference for A1 is higher" is expressed as A1 > A2. In the order of preference, A1 is superior (higher - ranked) to A2, A2 is inferior (lower - ranked) to A1, or it can also be expressed as A1 wins over A2 and A2 loses to A1.

[0005] Users list products from multiple online stores as candidates, carefully compare them, and purchase the product with the highest preference. At that time, when opening many browser tabs to compare candidate products, they tend to get confused and think "I don't know which product was good". Furthermore, many people feel inconveniences such as "it takes a long time", "dry eyes are painful", and "headaches" during product selection.

[0006] Patent Document 2 discloses a technology for calculating a value index of a product from a product information database to generate and display product comparison data. Patent Document 2 also discloses a technology for splitting the screen into two or four parts to select the specifications of a personal computer or the like and compare them simultaneously. Patent Document 3 discloses a technology for generating a product comparison information table from a product information table and displaying it on a split screen. However, none of them describe the technology of repeatedly comparing products, which is a feature of the present application, to select the optimal product.

[0007] Furthermore, to address the above issues, recommendation technologies have been developed that automatically suggest items that match the buyer's preferences and conditions using vector search and AI. However, ultimately, the buyer themselves needs to compare the candidate products and select one.

[0008] Candidate products are usually displayed in web browser tabs, and are also represented as bookmarks on the product pages. However, it is difficult to prioritize them or open them simultaneously, making comparison challenging. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] Japanese Patent Publication No. 2021-9616 Public Relations [Patent Document 2] Japanese Patent Publication No. 2002-63132 Public Relations [Patent Document 3] Japanese Patent Publication No. 2001-331680 Public Relations [Overview of the project] [Problems that the invention aims to solve]

[0010] The main challenge is to realize a technology that allows users to choose the product best suited to their purchase purpose simply by comparing two products, as shown in Figure 1, and selecting the one they prefer.

[0011] Buyers want to gather as many options as possible, compare the performance, design, and price of different products, and purchase the "number one" product that best suits their needs. However, it is difficult to compare numerous products simultaneously and choose the "number one" product in one go.

[0012] The comparator (comparison device) of the present invention is a technology that provides a mechanism for users to roughly select provisional candidates, such as number 1, number 2, or even number 3, simply by sliding their finger across the displayed products, and then ultimately select number 1 from among those candidates.

[0013] The number of top-ranked items to select for detailed comparison varies depending on the user and the product, so the system needs to be able to handle any number of selections. It's also necessary to allow users to stop the comparison process midway.

[0014] The most troublesome situation arises when the preferences for three or more products are in a stalemate. For example, A1 > A2 and A2 > A3, but it appears as if A3 > A1. Preventing such situations from occurring during the comparison process is also a challenge.

[0015] Furthermore, reducing the number of product comparisons and thus alleviating the burden on users, while simultaneously conserving CPU and storage resources, are also important challenges.

[0016] When making a selection, users may want to compare and choose from multiple product combinations, similar to coordinating outfits. The system also needs to accommodate situations where product advertisements are displayed in addition to the products the user has selected.

[0017] Furthermore, while the phrase "number 1, number 2, number 3" might bring to mind sorting problems, this invention is different. For example, if you select eight washing machines from various manufacturers from an online shop, you can sort by price, but you cannot quicksort by factors such as "cute color." Since users may stop comparing and move on to making a decision midway, this invention is an event-driven mechanism that responds to the user's intentions. [Means for solving the problem]

[0018] The present invention was made to solve the above problem, and is a comparator that allows a user who purchases products using a computer to perform an "n-number selection cycle" in which they select provisional 1st, 2nd, ... nth products with high preference from the product information of multiple purchase candidates (hereinafter referred to as products) displayed on the screen, and then compares the selected number of provisional 1st, 2nd, ... nth products to make a final selection. Assume that the process of the user selecting the products ranked provisionally first, second, …, nth in order of preference is called the “nth selection cycle”. A comparison display unit that displays two different said products to the user; A storage unit that stores the candidate list of the nth product, the preference of the comparison result between the products in the nth candidate list, and the product with the highest preference in the nth; A comparison preference unit that receives a preference event indicating that the user has selected one of the two products displayed by the comparison display unit; In the nth selection cycle, One or two products that have not been displayed yet are sequentially taken out from the nth candidate list, and the comparison display unit displays two different said products, When the comparison preference unit receives a preference event, the preference between the two products is stored in the storage means, and the product that has received the lower preference is not displayed again within the same selection cycle. When all the products in the candidate list of the nth candidate list have been displayed, under the condition that the law of transitivity between preferences holds, the product with the highest preference is stored in the storage means as the nth product. From the group of preference relationships between the products, the group of products that have not been directly compared with and preferred to the product with the highest preference is taken out and stored in the storage means as the next (n + 1)th candidate list, and the (n + 1)th selection cycle is executed, or the selection cycle is terminated. In the first provisional selection cycle, a comparison execution unit that stores the non-duplicate product group of the initial purchase candidates in the storage means as the first provisional candidate list and executes the first provisional selection cycle, A comparator characterized by comprising the above.

Effect of the Invention

[0019] An effect that the product optimal for the purpose of purchase can be selected only by comparing two products and choosing the preferred one. It is possible to select any number of products in order of their apparent quality from the entire set of candidate products, compare these candidates, and make a final selection. It also has the effect of preventing the preferences for three or more products from being in a state of indecision, omitting a large number of comparisons between products with undetermined preferences, reducing the number of product comparisons, alleviating the burden on the user, and simultaneously saving resources of the CPU and storage device.

[0020] More preferably, when a function of receiving an event for the user to start the next selection cycle is added, it also has the effect that even if the selection cycle is interrupted, the best product in that selection cycle can be selected.

Brief Description of the Drawings

[0021] [Figure 1] It is a diagram showing replacing products with lower preference while leaving products with higher preference. [Figure 2] It is a configuration diagram. [Figure 3] It is a list of products for purchase candidates in the initial state. [Figure 4] It is a state of prompting to compare two products side by side. [Figure 5] It is an overall flowchart. [Figure 6] It is a flowchart of the nth selection cycle. [Figure 7] It is the entirety of the first selection cycle. [Figure 8] It is the selection relationship of each selection in the first selection cycle. [Figure 9] It is a preference relationship diagram of the result of the first selection cycle. [Figure 10] It is a list of products for purchase candidates after the first selection cycle. [Figure 11] It is the entirety of the second selection cycle. [Figure 12] It is the selection relationship of each selection in the second selection cycle. [Figure 13] It is a preference relationship diagram of the result of the third selection cycle. [Figure 14]This is a preference relationship diagram when comparing [A8, A4, A6]. [Figure 15] This is a list of potential purchase items after the second selection cycle. [Figure 16] This is a comparison between number 1 and number 2. [Figure 17] This is a diagram showing the preference relationships for strategies that involve early termination. [Figure 18] This is a preference relationship diagram for strategies that result in a small residual value. [Figure 19] This is a diagram showing the selection of product combinations. [Figure 20] Here are some examples of noteworthy items. [Modes for carrying out the invention] [Examples]

[0022] (Claim 1, Claim 2, Claim 7, Claim 8) This embodiment describes the overall operation, components, and method of product comparison of the comparator, and provides a specific example of a user selecting a washing machine.

[0023] Figure 5 shows the overall flowchart of the comparator and product selection method. The comparison execution unit, described later, sequentially executes a selection cycle of provisional selections from number 1 to n shown in Figure 6, based on user instructions, and terminates the selection cycle through some means of termination (e.g., an exit button). After that, the user selects the final number 1 from the small number of candidates selected.

[0024] In Figure 5, the "candidate list" refers to a group of unique products that are compared within a single selection cycle.

[0025] Possible data structures include sets, lists ordered by addition or preference, and tables or records within a database. The advantages of using ordered lists instead of sets will be discussed later.

[0026] Figure 2 shows a diagram of the comparator configuration. These are implemented as blocks of a program running on hardware or the CPU.

[0027] The comparison display section, as shown in Figure 4, prompts the user to compare two products side by side. The two products displayed will not overlap with any previously displayed pair, although one of them may be a previously displayed product. Instead of a screen display, audio descriptions of the products can also be used.

[0028] In Figure 4, users compare product photos, videos, prices, product descriptions, reviews, discount coupons, etc., and choose the one they prefer. The comparison display can be shown in two ways: a "winner-take-all" method, where the selected product is kept and the unselected products are replaced with the next product, or a "tournament" method, where the winning products are displayed together.

[0029] The comparison operation unit receives preference events each time the user makes a selection and notifies the comparison execution unit. Possible selection methods include the user pressing a button (e.g., a heart icon) to select one of the products, the user sliding a less preferred product to display the next candidate as shown in Figure 1, the user tapping a more preferred product to swap it with the other product, the user tapping a less preferred product to swap it, and the user making a selection by voice.

[0030] Furthermore, when the user presses the button to start the next selection cycle, the comparison operation unit receives the next cycle start event and notifies the preference execution unit. The method of initiation can be any means, such as sliding, text input, or voice.

[0031] The memory unit stores a list of candidate products in the nth selection cycle, the preferences obtained as a result of comparing the products included in the nth candidate list, and the best-preference product that received the highest preference in the nth selection cycle.

[0032] Due to the large variety of data being handled, for the sake of brevity, when it is stated, for example, "Retrieve from the nth candidate list," it is implicitly assumed that data is read from the memory unit, and similarly, data is also written. In practice, this is implemented as input / output to a memory device, file device, or database.

[0033] Possible data structures for storing preferences like A1>A2 include using arrays or dictionaries to represent them in the form [A1, A2], representing them as directed graphs like A1<-A2 or A1->A2, and representing them as database records like (A1, 1), (A2, 2).

[0034] Figure 3 shows an example of an initial list of potential purchase items, representing multiple washing machines (A1, A2, A3, etc.) as products (product information). The format can be user-provided, retailer / manufacturer-provided, or browser / application-provided. At this stage, the order of preferences is irrelevant; products can be roughly categorized or arranged in any order. A1 and A2 are simply identifiers and have no relation to the order of preferences. Products can be arranged horizontally or vertically in the list.

[0035] The comparison execution unit performs the comparison operation using the above components.

[0036] Next, we will explain the technical features using the flowchart in Figure 6. The user initiates a comparison operation in some way, such as by tapping a button, using a gesture, or giving a voice command.

[0037] The comparison execution unit performs the following steps in one selection cycle. In the selection cycle in which option 1 is provisionally selected, the candidate list is made up of products obtained by removing duplicates from the initial product list (Figure 3). The candidate lists for the nth selection cycle, which are provided by the user, retailers and manufacturers, and browsers and applications, are created in the previous selection cycle (n-1).

[0038] Step 1: From the nth candidate list in the memory unit, retrieve one or two items that have not yet been displayed.

[0039] Step 2: As shown in Figure 4, display two different products in the comparison section and allow the user to compare them.

[0040] Step 3: As shown in Figure 1, when the user selects either product, the comparison preference unit receives a preference event. Upon receiving this notification, the preference relationship between the two products (e.g., A1 > A2) is stored in the memory unit.

[0041] Furthermore, the comparison display section ensures that products that receive a low preference are not displayed again within the same selection cycle. This is achieved by either removing the product from the nth candidate list or recording it in a separate array.

[0042] Step 4: Once all items on the candidate list have been displayed, the following rules are applied to determine the most preferred item, and it is stored in the storage device as "item number n". • Preferences between products that were not compared are unknown. However, if A1 > A2 and A2 > A3, then the transitive property A1 > A3 holds.

[0043] Step 5: If you want to terminate, stop the process at this point. If you want to perform a selection cycle, take out the group of products that were not preferred (lost) when directly compared to the best-preferenced product and store them as a candidate list for the next (n+1)th selection cycle. Return to Step 1 and start the (n+1)th selection cycle.

[0044] The following outlines the technical rationale for using "the group of products that were not preferred when directly compared to the nth most preferred product" as the candidate list for the next round. A verification using specific examples will be provided later.

[0045] This holds only when the following constraints are applied, and claims 1 and 7 are configured to include those constraints.

[0046] · For the compared products, there is no "tie", and a preference relationship of either A1 > A2 or A1 < A2 holds between A1 and A2. · A product that has received a low preference once is not redisplayed in the same selection cycle, so a loop of preference relationships that would result in a "three-way draw" does not occur. · A pair of two products that have been previously displayed is not displayed repeatedly. · If A1 > A2 holds in a previous comparison, A2 > A1 does not hold in a subsequent comparison (to prevent reversal of the preference relationship). · The preference between products that have not been compared is assumed to be unknown, but an indirect preference relationship (such as A1 > A3) holds by the law of transitivity.

[0047] Then, the result of one selection cycle becomes a "single" partially ordered set that satisfies the reflexive law, antisymmetric law, and transitive law. At this time, when an element A in the partially ordered set is covered by another element B, it means that A has a lower rank than B and there is no other element between them. By utilizing this property and making the group of products covered by the highest-preference product the next candidate list, it is possible to prevent unnecessary comparisons and contradictions in the preference relationship due to the inclusion of products with lower ranks, and to perform clear ranking.

[0048] Specifically, for the candidate list in the next cycle, use the covered elements, that is, the group of products (the losing group of products) that were directly compared with the highest-preference product and were not preferred. By this method, it is possible to maintain the consistency of the preference relationship while omitting unnecessary comparisons.

[0049] Also, when assuming large-scale simultaneous use by users, a significant saving effect on the CPU and memory area can be obtained.

[0050] Furthermore, in the configuration of claim 2 or claim 8, since the comparison preference unit executes step 4 immediately upon receiving the next cycle start event, the user can terminate the selection cycle at any time, enabling the determination of the next candidate list and early progression to the next cycle. This also results in the effect of further reducing the number of comparisons, as shown in Example 2.

[0051] If a feature is added that allows the user to go back and redo their selection any number of times, and the comparison results are canceled after that number of attempts, then a preference relationship loop cannot be created, and the above statement holds true.

[0052] Now, let's move on to a specific example. Figure 7 shows the selection cycle for choosing the first provisional item from the candidate list [Number 1], which includes washing machines as an element in Figure 3.

[0053] [Number 1] A1, A2, A3, A4, A5, A6, A7, A8

[0054] The top left cell in Figure 7 represents a comparison between A1 and A2. The next cell corresponds to Figure 1, illustrating how, following steps 1, 2, and 3, the comparison unit removes A2 from the comparison display unit because the user selected A1, and then moves on to comparing A1 and A3. The following dotted arrows show the comparison process from left to right, and then from left to right in the next row.

[0055] In Figure 7, the user ultimately left A5, so A5 is provisionally ranked number 1.

[0056] Figure 8 shows the comparison process written from top to bottom using preference notation. Since the number of comparisons for the number of products N in the candidate list is N-1, there are 8-1 = 7 preference relationships.

[0057] Based on the conditions in Step 5, preference relationships are linked, with highly preferred products at the top and less preferred products at the bottom, resulting in the preference relationship diagram (Hasse diagram) shown in Figure 9. Products that are not linked are not being compared.

[0058] Furthermore, organizing this preference relationship solely by rank results in the preference relationship equation [Equation 2]. The terms in parentheses represent groups of products with equivalent ranks for which the preference order has not yet been determined.

[0059] [Number 2] A5>[A6, A4, A7, A8]>A1>[A3, A2]

[0060] As mentioned in Step 5, in the next selection cycle to choose option 2, we only need to compare the set of items that follow A5, the highest preferred item, i.e., items whose preference is immediately after the highest preferred item, as shown in [Equation 3]. Items A1>[A3, A2] from the third item onward in [Equation 2] can be ignored (technically) in subsequent comparisons. A more detailed verification of this will be shown below.

[0061] [Number 3] A6, A4, A7, A8

[0062] Let's save [Mathematics 3] as a list of candidates for number 2 in our memory and move on to the selection cycle for number 2. An example is shown in Figure 11.

[0063] The comparison results are shown in Figure 12, and organizing them with Step 5 as the condition results in the preference relationship diagram in Figure 13. Furthermore, extracting only the ranks results in the preference relationship equation [Equation 5].

[0064] [Number 5] A7>[A4, A8]>A6

[0065] Looking closely at Figure 13, we see that there is no preference relationship between A8 and A6. In the next selection cycle to obtain option 3, should the candidate list be [A4, A8]? Or should we also compare A8 and A6, which have not yet been compared, and make the candidate list [A4, A8, A6]? If it's the latter, then the selection cycle 2 we've just gone through also lacks sufficient comparison.

[0066] Figure 13 shows the smallest model in which this question arises. We will examine all patterns before drawing a general conclusion.

[0067] When comparing [A4 and A8] to obtain the 3rd number, if the pattern is [number 6], the 3rd number is A8, and if the pattern is [number 7], the 3rd number is A4. The 3rd number is determined without any surprises.

[0068] [Number 6] A8>A4>A6

[0069] [Number 7] A4>[A6, A8]

[0070] When comparing [A8, A4, A6], there is no problem if A8 > A6, but problems arise if A6 > A8.

[0071] In that case, there are three possible patterns, including A6>A8: A6>A8>A4, A6>A4>A8, and A4>A6>A8. Of these, A4>A6 is already confirmed, so the case that contradicts it is excluded, leaving A4>A6>A8 in Figure 14. This pattern is included in [Number 7] A4>[A6, A8], and it can be seen that comparing only [A4, A8] is sufficient to resolve this question.

[0072] This minimal model appears similarly in various parts of the larger preference diagram. As we have examined here, combinations that have not been compared will ultimately lose to higher-ranking candidates, so we can conclude that only the group of products immediately following the top-ranking product should be included in the next candidate list.

[0073] The same effect is observed in Figure 9, where numerous comparisons between products with undetermined preferences can be omitted. This results in technical benefits such as reducing the effort required for comparisons by the user, as well as lowering CPU computation time, memory usage, and communication volume.

[0074] At this stage, provisional options 1 and 2 have been selected. The user can choose how many options to select. By creating a list of these n products as candidates, it is possible to use a comparator, as shown in Figure 16, when finally selecting option 1.

[0075] In Example 1, the user's actions are extremely simple: they simply choose their preferred option from each choice. This action effectively eliminates the complexity of comparing multiple candidates that we often experience in everyday life.

[0076] Furthermore, by continuously organizing the obtained preference relationships and constructing a data structure corresponding to the preference relationship diagram shown in Figure 9, it is also possible to visually display preference relationships to users in real time. [Examples]

[0077] (Claim 3, Claim 9) Next, we will explain techniques for further reducing the number of comparisons. Reduction means that the number of comparisons is smaller than the theoretical expected value N-1.

[0078] Even if we don't know exactly which products users prefer, we can make some inferences based on their past behavior and preferences.

[0079] For example, when a user adds a product to their initial list of candidates, they are unlikely to add a product that is clearly inferior to those already on the list. Therefore, it can be inferred that the product added later is more likely to be preferred. Methods can be considered to estimate the user's thinking based on various pieces of information, such as products that the user spent a long time selecting, products that they looked at for a long time, and products with high review ratings, which indicate a higher preference.

[0080] If a user terminates the comparison process midway, they receive a next cycle start event, immediately determining their best-preferenced product within that selection cycle. As a result, they can either proceed to the next selection cycle or make a final decision, reducing the number of comparisons compared to the expected value. In this "early termination strategy," it is effective to display products in order of user preference so that the user feels they are only seeing products they don't like and terminates the comparison.

[0081] For example, if products A1 through A6 are displayed in order of preference, and A1 beats A2, A3, and A4, and the user stops the selection cycle at that point, the preference relationship diagram shown in Figure 17 is formed. In this case, the number of comparisons is limited to 3, which is below the expected value of 5. [Examples]

[0082] (Claim 4, Claim 10) On the other hand, if you want to reduce the number of comparisons in the next selection cycle, the key is to reduce the number of products that lose to the top-preference product. Increasing the number of levels in the preference relationship diagram reduces the number of products that lose to the top-preference product, making the "sequential strategy," which displays products in order of least to most disliked by the user, effective.

[0083] For example, if products A1 through A6 are displayed in order of least preferred, comparisons will be repeated, such as A1 losing to A2, and A2 losing to A3. The comparison will continue until A6 is reached, forming the preference relationship diagram shown in Figure 18. In this case, the number of comparisons is N-1, as expected, but subsequent selection cycles become unnecessary.

[0084] In this way, by displaying products in order of general user preference (least preferred), it is possible to increase the number of steps in the preference relationship diagram and reduce the number of comparisons in the next selection cycle. If this order is set when creating the candidate list, the operation will be smoother.

[0085] Furthermore, even when user preferences are unknown in advance, the system can dynamically determine the next product to display using recommendation technologies such as vector search, LLM (Large-Scale Language Model), and product description similarity with each selection. For example, an early termination strategy prioritizes displaying products with high cosine similarity to the selected product, while a serial strategy prioritizes products with low cosine similarity.

[0086] Overall, an effective approach is to use a sequential strategy in the first selection cycle, displaying products in order of least preference to reduce the number of comparisons in subsequent cycles, and then using an early termination strategy in the second and subsequent selection cycles, displaying products in order of preference to further reduce the number of comparisons.

[0087] In this way, techniques that reduce the number of comparisons below the theoretical expected value significantly alleviate the burden on users performing comparison tasks, and also enable more efficient use of CPU time and memory. [Examples]

[0088] (Claim 5) One challenge is the need to compare and select from multiple product combinations. Figure 19 shows an example of selecting a tank top and skirt outfit. In this example, two outfits are compared where the skirts are the same, but the tank tops are different. Possible approaches include swapping out entire product combinations, or swapping out only some of the items (tank tops or skirts) within a combination.

[0089] Automatically generating outfits from tank tops and skirts in a candidate list not only allows users to compare outfits but also eliminates the need for users to create combinations themselves. Furthermore, since images from apparel shops often include people and backgrounds, using object detection techniques to extract and display skirts and other elements is also an effective method. [Examples]

[0090] (Claim 6) In addition to the products users have gathered, there is also the challenge of displaying product recommendations and advertisements. By recommending products and displaying advertisements based on user preference history, in addition to the products in the candidate list, the convenience for users is improved, and advertisers gain opportunities to sell their products.

[0091] If a user selects a recommended product, the system can process this by adding more products to the candidate list and applying the method described above. [Examples]

[0092] (Claim 11) As shown in Figure 20, a comparison display that compares two products based on user-centric comparison items such as price, power consumption, and review ratings, and shows which is superior, is also an effective format. For example, if A5 shows that price is superior and A7 shows that power consumption is superior, users can judge the superiority or inferiority at a glance. In addition to superiority or inferiority, analog values ​​such as price can be represented with a bar graph to facilitate comparison. Comparison items can be specified by the user, selected by AI, or pre-defined for each product category. [Examples]

[0093] (Claim 12) Streaming the comparison display as a video is also an effective method. This allows many people to refer to the selection results when making a purchase. [Examples]

[0094] At the end of the selection cycle, there is a challenge in making the results easier to understand by sorting the list of potential purchase items by preference. Figure 10 shows the list of potential purchase items after the first selection cycle, and Figure 15 shows the list of potential purchase items after the second selection cycle. In practice, the items can be sorted based on the preference relationship equation or preference relationship diagram at the end of the selection cycle, or the items in the product list can be moved each time a comparison is made. This results in the list of potential purchase items being sorted by preference, making the results easier to understand. [Examples]

[0095] If family or colleagues are not convinced during the selection process, it can lead to serious problems after the purchase, such as "getting yelled at by my wife." This can be resolved by having the communication unit in Figure 2, in response to preference events, send and receive the comparison operation and the post-comparison preference relationships to the comparators of family and colleagues, thus synchronizing them. This has the effect of reducing the likelihood of disputes after the purchase. Using it in conjunction with chat or video conferencing is also effective.

[0096] When decisions are made through voting, a positive effect can be achieved even with a large number of participants. [Examples]

[0097] There is also a demand for resuming selection from where it left off after the selection process has been terminated. By storing data equivalent to the preference relationship diagram in the memory unit when the selection is terminated, it is possible to restore the product display when the selection is resumed. [Examples]

[0098] Sometimes it can be difficult to decide between A1 and A2. Processing product information and images with AI to display advice or help users choose one over the other can be effective. If data corresponding to the relationships between each choice or a diagram of those relationships is also sent, the AI ​​can understand the user's preference trends and provide more appropriate advice. [Examples]

[0099] If you're unsure whether A1 or A2 is better, you can use a random number generator to make the selection. Displaying a "ladder lottery" and letting the user choose their starting point has the effect of allowing them to make a choice without appearing to be playing catch-up. [Examples]

[0100] By collecting selection relationships on the server, it becomes possible to recommend or display advertisements to users for sellers and products with fewer selections, or for best-selling sellers and products. This has the effect of broadly distributing revenue opportunities among sellers and producers. [Examples]

[0101] If a user doesn't like any of the products, they can generate and publish a request for a specific product based on their selection history. This allows them to receive suggestions for products that match their preferences from sellers who receive their requests.

Claims

1. A comparator used by computer users to make a purchase of goods, which performs an "n-number selection cycle" in which they select provisional items 1st, 2nd, ... nth with the highest preference from the information of multiple purchase candidates (hereinafter referred to as "goods") displayed on the screen, and then compares the selected number of provisional items 1st, 2nd, ... nth to make a final selection. If we call the process of selecting the aforementioned products in order of preference, from 1st to 2nd...nth, the "nth selection cycle," then, A comparison display unit that displays two different products to the user, A memory unit that stores a list of candidate products for item n, preferences resulting from comparing the products on the candidate list for item n, and the highest preferred item for item n. A comparison preference unit receives a preference event indicating that the user has selected one of the two products displayed by the comparison display unit, In the aforementioned nth selection cycle, Select one or two products from the aforementioned nth candidate list that have not yet been displayed, and display two different products on the comparison display unit. When the comparison preference unit receives a preference event, it stores the preference between the two products in the storage means, and prevents the product that received a low preference from being displayed again within the same selection cycle. When all the products in the nth candidate list have been displayed, the product that received the highest preference under conditions where the transitivity law between preferences holds is stored in the storage means as the nth product. From the group of preference relationships between the aforementioned products, the group of products that were not preferred when directly compared with the most preferred product is extracted and stored in the storage means as the next (n+1) candidate list, and the (n+1) selection cycle is executed, or the selection cycle is terminated. In the provisional selection cycle No. 1, the comparison execution unit stores the initial group of non-overlapping purchase candidates as the provisional candidate list No. 1 in the storage means and executes the provisional selection cycle No.

1. A comparator equipped with the following features.

2. The comparison execution unit, The comparator according to claim 1, characterized in that, when all the products in the candidate list of nth candidate list are displayed, or when the comparison preference unit receives the next cycle start event, the best-preference product that received the highest preference under conditions where a transitivity law holds between preferences is stored in the storage means as the nth product.

3. The comparator according to claim 1 or 2, characterized in that the comparison execution unit displays the products on the comparison display unit in an order that is presumed to be highly preferred by the user, terminates the selection cycle when it receives the next cycle start event, and determines the nth product while there are still products remaining in the candidate list, thereby reducing the number of comparisons.

4. The comparator according to claim 1 or 2, characterized in that the comparison execution unit displays the products on the comparison display unit in an order in which the user's preference is presumed to be low, thereby reducing the number of product groups that are not preferred when directly compared with product number n, and reducing the number of comparisons in the selection cycle of product number n+1.

5. The comparator according to claims 1 to 4, further comprising a comparison display unit that displays two sets of multiple products, and a comparison execution unit that swaps the sets of products or swaps only one product in the set of products based on a preference event in the comparison operation unit.

6. The comparator according to claims 1 to 5, which displays recommended products or advertisements in the comparison display section according to the history of preference relationships.

7. A product selection method used by a computer-based buyer to make a purchase, in which the user selects tentatively preferred products (1st, 2nd, ... nth) from a list of product information (hereinafter referred to as "products") displayed on the screen, performs an "n-number selection cycle," compares the selected few tentative products (1st, 2nd, ... nth) to make a final selection. In the nth selection cycle, the comparison execution unit Step 1 involves sequentially selecting one or two items from the aforementioned nth candidate list that have not yet been displayed, Step 2 involves displaying two different products in the comparison display section, Step 3: When the comparison preference unit receives a preference event indicating that the user has selected one of the products, it stores the preference between the two products in the storage means, and ensures that the product that received a lower preference is not displayed again within the same selection cycle. Step 4 involves displaying all the products on the nth candidate list, and then storing the product that received the highest preference under conditions where the transitivity law between preferences holds as the nth product in the storage means. The system includes step 5, which involves either selecting the group of products that were not selected when directly compared with the most preferred product from the group of preferences among the products, storing them in the storage means as the next (n+1) candidate list, and executing the (n+1) selection cycle, or ending the selection cycle. Step 0: The comparison execution unit stores the initial group of non-duplicate purchase candidates in the storage means as the provisional candidate list No. 1 and executes the provisional selection cycle No.

1. A product selection method that includes the following features.

8. The comparison execution unit also, The product selection method according to claim 7, further comprising step 4, in which, at the time when all products in the nth candidate list are displayed, or when a next cycle start event is received to start the next selection cycle, the most preferred product, which has received the highest preference under conditions where a transitivity law holds between preferences, is stored in the storage means as the nth product.

9. The product selection method according to claims 7 to 8, characterized in that, in the step of the comparison execution unit taking out a product from the candidate list and in the step of displaying the product on the comparison display unit, the products are displayed on the comparison display unit in an order that is presumed to be highly preferred by the user, and in the step of receiving the next cycle start event, the nth product is determined while there are still products remaining in the candidate list, thereby reducing the number of comparisons.

10. The product selection method according to claims 7 to 8, characterized in that, in the step of the comparison execution unit taking out a product from the candidate list and the step of displaying the product on the comparison display unit, the product is displayed on the comparison display unit in an order in which the user's preference is presumed to be low, thereby reducing the number of product groups that are not preferred when directly compared with product number n, and reducing the number of comparisons in the n+1 selection cycle.

11. The comparator according to claims 1 to 6, characterized in that the comparison display unit displays which of the two products is superior with respect to the comparison item of interest to the user.

12. The comparator according to claims 1 to 6 to 11, characterized in that the display of the comparison display unit is streamed as a video.

13. A program for running the comparator described in claims 1 to 12.

Images