HK1192035B - Method for providing search result, and apparatus thereof - Google Patents

Description

Method and device for providing search results

Technical Field

The present application relates to the field of communication technology, and in particular to a method and device for providing search results.

Background Art

Currently, most shopping websites offer users a search function that allows them to search for product information of desired products. Specifically, the user sends a search term to the shopping website's server. The server searches for product information related to the received search term based on the received search term and provides the searched product information to the user.

In actual applications, users are generally more interested in product information that is ranked higher. Therefore, after searching for product information, the server needs to sort the searched product information and provide the sorted product information to the user. Since users are generally more interested in the prices included in the product information, in the prior art, the server can sort the searched product information in descending or descending order based on the prices included.

However, different sellers may set different prices for the same product. For example, a buyer might search for a certain brand of running shoes. Among the product information retrieved by the server, 10 products have a price of 20 yuan, 20 have a price of 100 yuan, and 5 have a price of 300 yuan. However, the actual price of the brand of running shoes should be 120 yuan. Obviously, the buyer would intuitively think that 100 yuan is more reasonable, and thus pay attention to the 20 products with a price of 100 yuan.

However, in the prior art, the server can only sort the searched product information in order of the prices contained, from high to low or from low to high. Assuming that the server sorts the searched product information in the above example in order of prices from low to high, the sorting result is: 10 product information containing a price of 20 yuan, 20 product information containing a price of 100 yuan, and 5 product information containing a price of 300 yuan.

Obviously, the server did not rank the 20 products priced at 100 yuan at the front, but instead ranked the 10 products priced at 20 yuan at the front. Therefore, after the server provides the sorted product information to the user, the products corresponding to the product information at the front are often not the products that the user thinks are reasonably priced. As a result, the user will constantly set filtering conditions and re-search, which increases the pressure on the server.

Summary of the Invention

The embodiments of the present application provide a method and apparatus for providing search results, which is used to solve the problem in the prior art that after the server provides sorted product information to the user, the user still needs to continuously set filtering conditions to re-search, causing great pressure on the server.

An embodiment of the present application provides a method for providing search results, including:

The server searches for relevant product information based on the received search term; and

Finding a predetermined optimal price corresponding to the search term; and

The absolute value of the difference between the price contained in each searched product information and the found optimal price is determined respectively, and the searched product information is sorted and provided in ascending order according to the respectively determined absolute values.

An embodiment of the present application provides a device for providing search results, including:

A search module, used to search for relevant product information based on the received search terms;

An optimal value determination module, configured to find a predetermined optimal price value corresponding to the search term;

A module is provided for respectively determining the absolute value of the difference between the price contained in the searched product information and the found optimal price value, and sorting and providing the searched product information in ascending order according to the respectively determined absolute values.

An embodiment of the present application provides a method and device for providing search results. When the server performs a search based on a received search term, the method searches for a predetermined optimal price value corresponding to the search term, determines the absolute value of the difference between the price contained in each searched product information and the optimal price value, and sorts and provides the searched product information in order of the determined absolute values from small to large. Through the above method, among the product information provided by the server to the user, the product information with the highest sorting order is the product information whose contained price has the smallest difference with the optimal price value. Compared with the method in the prior art that provides search results in a single order from high to low or from low to high according to the contained price, the method can effectively reduce the number of times the user sets the filter conditions and searches again, thereby reducing the pressure on the server.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a process for providing search results according to an embodiment of the present application;

FIG2 is a process of fitting the probability density function of the price contained in the product information searched based on the search term provided by an embodiment of the present application;

FIG3 is a schematic diagram of the structure of a device for providing search results provided in an embodiment of the present application.

DETAILED DESCRIPTION

Since the server in the prior art can only sort and provide the searched product information in order of the prices contained therein from high to low or from low to high, the prices contained in the top-ranked product information provided by the server in the prior art are often far from what most users consider to be reasonable prices. This will cause users to constantly set filtering conditions and search again, which increases the pressure on the server.

In the embodiment of the present application, the server predetermines the optimal price value corresponding to the search term, which is the price that most users consider to be reasonable among the prices contained in the product information that can be searched based on the search term. When searching based on the search term, the searched product information is sorted and provided in such a manner that the smaller the absolute value of the difference between the price contained in the searched product information and the predetermined optimal price value corresponding to the search term, the higher the ranking. The product information with a higher ranking provided to the user is the product information whose price is closer to the optimal price value corresponding to the search term, so as to reduce the number of times the user re-searches by setting filtering conditions and alleviate the pressure on the server.

The embodiments of the present application are described in detail below with reference to the accompanying drawings.

FIG1 is a process for providing search results according to an embodiment of the present application, which specifically includes the following steps:

S101: The server searches for relevant product information based on the received search term.

Similar to the prior art, when a user wants to search for product information of a certain product, he sends the corresponding search term to the server. After the server receives the search term sent by the user, it searches for product information related to the search term based on the received search term.

S102: Finding a predetermined optimal price value corresponding to the search term.

In an embodiment of the present application, the server has previously determined a corresponding optimal price value for each search term recorded in the search log. Therefore, after the server performs a search based on the received search term, it searches for the predetermined optimal price value corresponding to the search term.

The method for predetermining the optimal price value corresponding to the search term may be: taking the average of the prices of all product information that can be searched based on the search term as the optimal price value corresponding to the search term, or taking the median of the prices of all product information as the optimal price value corresponding to the search term. Of course, the optimal price value corresponding to the search term may also be manually set.

S103: Determine the absolute value of the difference between the price included in each searched product information and the found optimal price, sort the searched product information in ascending order of the determined absolute values, and provide the results.

That is, when sorting the searched product information, the smaller the absolute value of the difference between the included price and the optimal price value, the higher the ranking. Since the predetermined optimal price value corresponding to the search term is closer to what most users consider to be a reasonable price for the product searched with the search term, after the searched product information is sorted according to the sorting method in step S103 and provided to the user, the number of times the user sets the filtering conditions to re-search can be effectively reduced, thereby reducing the pressure on the server.

In an embodiment of the present application, the server needs to determine the corresponding optimal price value for each used search term recorded in the search log in advance. The method for determining the optimal price value corresponding to a certain search term is specifically that the server searches the search log for product information that meets the specified conditions in the product information searched for each time a search was performed based on the search term in the past, and determines the optimal price value corresponding to the search term based on the price contained in the product information found. Among them, the product information that meets the specified conditions includes the product information clicked by the user. Of course, the product information that meets the specified conditions can also include product information collected by the user, product information for which the corresponding product has been ordered, product information shared, and product information for which other users have performed specified operations. In addition, when the server searches for product information that meets the specified conditions that was searched for each time based on the search term in the past, it can search for product information that meets the specified conditions in the product information that was searched for each time a search was performed based on the search term within a set time period in the past.

In actual applications, after the server provides the user with product information found based on the search term, the user will often focus on product information that they believe has a reasonable price. For those product information that they believe has a reasonable price, the user will usually click on it to view the product information in detail. However, for product information that they believe has an unreasonable price, the user will usually not click on it and will ignore it. Therefore, the server determines the optimal price corresponding to the search term based on the prices of the product information that was clicked in each of the product information searches based on the search term.

For example, assuming that the optimal price corresponding to the search term "1g graphics card" is to be determined, the server searches the search log for past searches using the search term "1g graphics card." Assuming that three searches are found, the server further searches for product information that meets the specified conditions among the product information found during each of the three searches. In other words, the server searches for product information that was clicked by the user after each search using the search term "1g graphics card." Assuming that the product information clicked in the first search is product A, the product information clicked in the second search is product B, and the product information clicked in the third search is product C, the server determines the optimal price corresponding to the search term "1g graphics card" based on the prices contained in the found information about product A, the prices contained in the found information about product B, and the prices contained in the found information about product C. Among them, the average value of the prices contained in the found product information can be determined as the optimal price value corresponding to the search term, that is, the average value of the prices contained in the information of product A, product B, and product C respectively can be used as the optimal price value corresponding to the search term.

Of course, when determining the optimal price corresponding to the search term based on the prices contained in the found product information, the prices contained in the found product information can be converted into a unified metric price according to actual circumstances. For example, the prices contained in the found product information can be converted into the price per product, or the price per kilogram, or the price per liter, or the price per square meter, etc. according to actual circumstances.

Through the above method, the server can more accurately determine the optimal price value corresponding to the search term, so that when subsequent searches are performed based on the search term, after sorting and providing the searched product information through step S103 shown in Figure 1, the number of times the user sets filtering conditions and re-searches can be further reduced, thereby further reducing the pressure on the server.

Considering that there are many similar search terms in actual applications, such as "1g graphics card" and "graphics card 1g", the product information found when searching with these two search terms is basically the same. Therefore, the server can organize the search terms recorded in the search log and combine similar search terms into a normalized search term. For example, the search terms "1g graphics card" and "graphics card 1g" can be combined into the normalized search term "1g graphics card". When searching for product information that meets the specified conditions among the product information found in previous searches using this search term, the search log is searched based on the normalized search term of this search term.

For example, after integrating the search terms "1g graphics card" and "graphics card 1g" recorded in the search log into the normalized search term "1g graphics card", if you want to determine the optimal price value corresponding to the search term "graphics card 1g", then determine the normalized search term of the search term "graphics card 1g" as "1g graphics card", and search the search log for the product information clicked by the user each time the search was based on "1g graphics card" in the past, and determine the optimal price value corresponding to the search term "graphics card 1g" based on the price contained in the found product information.

Preferably, since in actual applications, for product information searched with the same search term, the probability density distribution of the prices contained in each searched product information roughly conforms to the mixed double Gaussian probability distribution, therefore, when the server determines the optimal price value corresponding to the search term based on the prices contained in each product information found that meets the specified conditions, it can use the mixed double Gaussian model to fit the probability density function of the prices contained in the product information searched based on the search term based on the prices contained in the found product information, and determine the price corresponding to the maximum probability density as the optimal price value corresponding to the search term based on the fitted probability density function.

Among them, the process of using the mixed double Gaussian model to fit the probability density function of the price contained in the product information searched based on the search term is specifically shown in Figure 2. Figure 2 is a process of fitting the probability density function of the price contained in the product information searched based on the search term provided by an embodiment of the present application, which specifically includes the following steps:

S201: Determine a total average value of the prices included in the found product information based on the prices included in the found product information.

For example, to determine the optimal price for the search term "1g graphics card", the server searches the search log for past searches with the search term "1g graphics card" and finds product information that meets the specified criteria. Assuming that n product information that meets the specified criteria is found, and the price contained in the i-th product information is denoted as x _i , then determine the total average price of these n found product information.

S202: Determine a total standard deviation of the prices included in the found product information based on the prices included in the found product information and the determined total average value.

Continuing with the above example, after determining the total average of the prices contained in the n product information, the formula is used to determine the total standard deviation σ of the prices contained in these n product information.

Preferably, for the prices contained in the found product information, if the price contained in a certain product information differs significantly from the determined overall average, then the price contained in the product information is abnormal data, and abnormal data will affect the accuracy of subsequent fitting. Therefore, in the embodiment of the present application, the server can also remove abnormal data from the prices contained in each found product information. Specifically, continuing with the above example, for the price x i contained in the i-th product information among the n found product information, if the value is greater than 3, then the price x _i contained in the i-th product information is determined to be abnormal data, and the price x _i contained in the i-th product information is removed. Alternatively, if the value is not within a set range, then the price x _i contained in the i-th product information is determined to be abnormal data, and the price x _i contained in the i-th product information is removed. The set range can be _, for example, 0.2 to 1.8.

S203: taking the product of the determined total average value and a first setting parameter as a first average value, and taking the product of the determined total average value and a second setting parameter as a second average value, wherein the first setting parameter is greater than 1 and the second setting parameter is less than 1.

Continuing with the above example, the first average value is determined and the second average value is determined, where a is the first setting parameter and b is the second setting parameter. In the embodiment of the present application, it is sufficient to ensure that one of a and b is greater than 1 and the other is less than 1, for example, a is 0.3 and b is 1.3.

S204: The determined total standard deviation is used as both the first standard deviation and the second standard deviation.

Continuing with the above example, the first standard deviation σ ₁ is determined to be σ, and the second standard deviation σ ₂ is determined to be σ.

S205: For each price contained in the product information found, use a formula to determine the median value corresponding to each price.

Where p is the initial weight value, which can be set to 0.5.

Continuing with the above example, for the price x _i contained in the i-th product information among the n product information found, the above formula is used to determine the median value gamma(i) corresponding to the price x _i contained in the i-th product information.

S206: Re-determine the first average value using a formula according to the determined median value corresponding to each price, and re-determine the second average value using a formula.

S207: Based on the newly determined first and second mean values, use the formula to re-determine the first standard deviation and the formula to re-determine the second standard deviation.

Wherein, mu ₁ and mu ₂ in the formula are the first average value and the second average value re-determined in step S206.

S208: Re-determine the weight using a formula.

Where n is the number of product information found.

S209: Determine whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions. If so, execute step S211; otherwise, execute step S210.

In an embodiment of the present application, the method for determining whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions is specifically as follows: determining whether the difference between the re-determined first average value and the first average value determined last time is within the set range, determining whether the difference between the re-determined second average value and the second average value determined last time is within the set range, determining whether the difference between the re-determined first standard deviation and the first standard deviation determined last time is within the set range, and determining whether the difference between the re-determined second standard deviation and the second standard deviation determined last time is within the set range; when at least one of the above four judgments is yes, determining whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions.

Of course, when the results of at least two or all of the above four judgments are yes, it is also possible to judge whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions.

S210: Re-determine the median value corresponding to each price based on the re-determined first average value, second average value, first standard deviation, second standard deviation, and weight, and return to step S206.

That is, based on the newly determined first average, second average, first standard deviation, second standard deviation, and weight, the formula is used to re-determine the median price for each product found. Based on each newly determined median, the first average, second average, first standard deviation, second standard deviation, and weight are then determined until the determined first average, second average, first standard deviation, and second standard deviation meet the specified conditions.

S211: Based on the newly determined first mean, second mean, first standard deviation, second standard deviation, and weight, a function is determined as a fitted probability density function of the prices contained in the product information found based on the search term.

Continuing with the above example, when the first mean, second mean, first standard deviation, and second standard deviation satisfy the set conditions, the function obtained based on the first mean, second mean, first standard deviation, and second standard deviation, as well as the final weights, is the fitted probability density function of the prices contained in the product information found based on the search term "1g graphics card."

In the subsequent process, the price corresponding to the maximum probability density can be determined as the optimal price value corresponding to the search term "1g graphics card". That is, the value of the independent variable x corresponding to the maximum function value f(x) _max of the above function is the optimal price value corresponding to the search term "1g graphics card". This is because the above probability density function is determined based on the prices contained in the product information found that meets the specified conditions, and the product information that meets the specified conditions is the product information clicked by the user. Therefore, the above probability density function reflects the probability of users clicking on product information containing different prices in the searched product information when searching with the search term "1g graphics card". The greater the probability of clicking on product information containing a certain price, the more users believe that the price is a reasonable price. Therefore, the price corresponding to the maximum probability density is used as the optimal price value corresponding to the search term. This can more accurately determine the optimal price value corresponding to the search term, further reducing the number of times users set filtering conditions and re-search, and reducing the pressure on the server.

After obtaining the fitted probability density function of the prices contained in the product information searched based on the search term, since the maximum value of the function must appear between mu ₁ and mu ₂ , the method for determining the value of the independent variable x corresponding to when the function takes the maximum function value f(x) _max can be: dividing the range of mu ₁ to mu ₂ into several sub-ranges. For example, the range of mu ₁ to mu ₂ can be divided into n sub-ranges, namely mu ₁ to mu ₁ +d, mu ₁ +d to mu ₁ +2d, mu ₁ +2d to mu ₁ +3d, ..., mu ₁ +(n-1)d to mu ₂ , wherein a value of x is randomly selected in each divided sub-range, and the function value f(x) corresponding to each selected value of x is determined according to the obtained probability density function. The value of x corresponding to the largest function value among each determined function value f(x) is determined as the value of the independent variable x corresponding to when the function takes the maximum function value f(x) _max , that is, the optimal price function corresponding to the search term.

Preferably, since in statistics, if one wants to determine whether the probability density distribution of certain samples conforms to a mixed double Gaussian probability distribution, the number of different samples is required to be greater than a certain threshold. Therefore, in the embodiment of the present application, before using the process shown in FIG2 to fit the probability density function of the prices contained in the product information searched based on the search term, it is also necessary to determine that the number of product information found that meets the specified conditions and contains different prices is greater than the set threshold.

In addition, considering that the prices contained in the product information searched based on certain search terms are often high in actual applications, the amount of calculation required when fitting the probability density function of the prices contained in the product information searched based on the search terms using the process shown in Figure 2 is large. Therefore, in the above step S201, the price contained in each product information found can be logarithmically operated, and the result obtained can be used as the price contained in each product information.

For example, for the price x _i contained in the i-th product information among the n products that meet the specified conditions, the value of log ₂ x _i can be re-used as the price contained in the i-th product information. Subsequently, the fitting process shown in Figure 2 is performed based on the re-determined prices contained in each product information. Assuming that the optimal price value corresponding to the search term is ultimately determined to be X according to the process shown in Figure 2, 2 ^X is also re-determined as the optimal price value corresponding to the search term.

After obtaining the optimal price value corresponding to the search term, the obtained optimal price value of the search term can also be manually adjusted, and the optimal price value of the search term can be saved to find the optimal price value of the search term when performing a subsequent search based on the search term, and the searched products can be sorted and provided according to the prices contained in the searched product information and the optimal price value corresponding to the search term. Of course, in the embodiment of the present application, when the server sorts the searched product information, in addition to sorting according to the prices contained in the searched product information and the optimal price value corresponding to the search term, it can also sort the searched product information based on factors such as the relevance of the searched product information to the search term, the quality of the product corresponding to the searched product information, and the quality of the seller corresponding to the searched product information.

The server may also sort and provide the searched product information based on the prices contained in each searched product information and the optimal price value corresponding to the search term only when it receives a request from the user to sort by the optimal price value. If no request to sort by the optimal price value is received, the searched product information will be sorted and provided according to the default sorting rules.

FIG3 is a schematic diagram of a structure of a device for providing search results according to an embodiment of the present application, which specifically includes:

Search module 301, for searching for relevant product information based on received search terms;

The optimal value determination module 302 is used to find the predetermined optimal price value corresponding to the search term;

The providing module 303 is used to respectively determine the absolute value of the difference between the price contained in each searched product information and the found optimal price value, and sort and provide the searched product information in ascending order according to the respectively determined absolute values.

The optimal value determination module 302 is specifically used to search for product information that meets specified conditions in each past search based on the search term, as recorded in the search log, wherein the product information that meets the specified conditions includes product information clicked by the user; and determine the optimal price value corresponding to the search term based on the price contained in the found product information.

The optimal value determination module 302 is specifically used to use a mixed double Gaussian model to fit the probability density function of the prices contained in the product information searched based on the search term, and determine the price corresponding to the maximum probability density as the optimal price value corresponding to the search term based on the fitted probability density function.

The optimal value determination module 302 specifically includes:

Initialization unit 3021 is configured to determine, based on the prices included in the found product information, a grand average of the prices included in the found product information; determine, based on the prices included in the found product information and the determined grand average, a grand standard deviation of the prices included in the found product information; multiply the determined grand average by a first setting parameter as a first grand average; and multiply the determined grand average by a second setting parameter as a second grand average, wherein the first setting parameter is greater than 1 and the second setting parameter is less than 1; and use the determined grand standard deviation as both the first standard deviation and the second standard deviation;

The training unit 3022 is used to determine the median value corresponding to each price contained in each product information found using the formula, where p is the initial weight, mu ₁ is the first mean value, mu ₂ is the second mean value, σ ₁ is the first standard deviation, σ ₂ is the second standard deviation, and x _i is the price contained in the i-th product information found, gamma(i) is the median value corresponding to the price contained in the i-th product information found; based on the determined median value corresponding to each price, the first average value is re-determined using the formula, and the second average value is re-determined using the formula; based on the re-determined first average value and second average value, the first standard deviation is re-determined using the formula, and the second standard deviation is re-determined using the formula; the weight is re-determined using the formula, wherein n is the number of product information found; it is judged whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions; when the judgment result is no, the median value corresponding to each price is re-determined based on the re-determined first average value, second average value, first standard deviation, second standard deviation, and weight, and the first average value, second average value, first standard deviation, second standard deviation, and weight are continued to be determined based on each re-determined median value until the determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions;

Function determination unit 3023 is configured to, when the training unit 3022 determines that the newly determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions, determine a function based on the newly determined first average value, second average value, first standard deviation, second standard deviation, and weights as a fitted probability density function of the prices contained in the product information retrieved based on the search term.

The training unit 3022 is specifically used to determine whether the re-determined first average value, the second average value, the first standard deviation, and the second standard deviation meet the set conditions when the difference between the re-determined first average value and the last determined first average value is within the set range, or when the difference between the re-determined second average value and the last determined second average value is within the set range, or when the difference between the re-determined first standard deviation and the last determined first standard deviation is within the set range, or when the difference between the re-determined second standard deviation and the last determined second standard deviation is within the set range.

An embodiment of the present application provides a method and device for providing search results. When the server performs a search based on a received search term, the method searches for a predetermined optimal price value corresponding to the search term, determines the absolute value of the difference between the price contained in each searched product information and the optimal price value, and sorts and provides the searched product information in order of the determined absolute values from small to large. Through the above method, among the product information provided by the server to the user, the product information with the highest sorting order is the product information whose contained price has the smallest difference with the optimal price value. Compared with the method in the prior art that provides search results in a single order from high to low or from low to high according to the contained price, the method can effectively reduce the number of times the user sets the filter conditions and searches again, thereby reducing the pressure on the server.

Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims (10)

1. A method for providing search results, characterized in that it includes: The server searches for relevant product information based on the received search terms, including product prices; and From the pre-determined optimal price values for each search term recorded in the search log, the optimal price value corresponding to the received search term is retrieved, wherein the optimal price value corresponding to a search term is determined by looking up the prices of product information that meets specified conditions and were found in past searches based on the search term, as recorded in the search log; and Determine the absolute value of the difference between the price contained in each searched product information and the optimal price found, and sort and provide the searched product information in ascending order of the determined absolute values; The optimal price corresponding to the pre-determined search term is the price that the user considers reasonable for the products found using that search term. 2. The method as described in claim 1, characterized in that determining the optimal price value corresponding to the search term specifically includes: The server searches the search log for product information that meets specified conditions from the product information found during each past search based on the search term. The product information that meets the specified conditions includes product information that has been clicked by the user. Based on the prices contained in the found product information, determine the optimal price value corresponding to the search term. 3. The method as described in claim 2, characterized in that determining the optimal price value corresponding to the search term based on the price contained in the found product information specifically includes: Based on the prices contained in the found product information, a mixture of Gaussian models is used to fit the probability density function of the prices contained in the product information searched based on the search terms; Based on the fitted probability density function, the price corresponding to the maximum probability density is determined as the optimal price value for the search term. 4. The method as described in claim 3, characterized in that, fitting the probability density function of the price contained in the product information searched based on the search term using a mixture of Gaussian models, specifically including: Based on the prices contained in the found product information, determine the overall average of the prices contained in the found product information; Based on the prices contained in the found product information and the determined total average, determine the total standard deviation of the prices contained in the found product information; The product of the determined total average value and the first set parameter is used as the first average value, and the product of the determined total average value and the second set parameter is used as the second average value, wherein the first set parameter is greater than 1 and the second set parameter is less than 1; The determined total standard deviation is used as both the first and second standard deviations. For each product information found, the price contained in the product information is determined by a formula, where p is the initially set weight, mu ₁ is the first average value, mu ₂ is the second average value, σ ₁ is the first standard deviation, σ ₂ is the second standard deviation, _xi is the price contained in the i-th product information found, and gamma(i) is the median value corresponding to the price contained in the i-th product information found. Based on the median value corresponding to each determined price, the first average value is redetermined using a formula, and the second average value is redetermined using a formula. Based on the newly determined first and second averages, the first standard deviation is redetermined using the formula, and the second standard deviation is redetermined using the formula. The weights are redefined using a formula, where n is the number of product information items found. Determine whether the newly determined first mean, second mean, first standard deviation, and second standard deviation meet the set conditions; If so, then based on the newly determined first average, second average, first standard deviation, second standard deviation, and weights, a function is determined as the probability density function of the price contained in the product information searched based on the search term. Otherwise, based on the newly determined first average, second average, first standard deviation, second standard deviation, and weight, the median value corresponding to each price is re-determined, and based on each newly determined median value, the first average, second average, first standard deviation, second standard deviation, and weight are continued to be determined until the determined first average, second average, first standard deviation, and second standard deviation satisfy the set conditions. 5. The method as described in claim 4, characterized in that determining whether the newly determined first average, second average, first standard deviation, and second standard deviation satisfy the set conditions specifically includes: Determine if the difference between the newly determined first average and the previously determined first average is within a set range; or Determine if the difference between the newly determined second average and the previously determined second average is within a set range; or Determine if the difference between the newly determined first standard deviation and the previously determined first standard deviation is within the set range; or Determine if the difference between the newly determined second standard deviation and the previously determined second standard deviation is within the set range. 6. An apparatus for providing search results, characterized in that it comprises: The search module is used to search for relevant product information based on received search terms, the product information including the price of the product; The optimal value determination module is used to find the optimal price value corresponding to the received search term from the corresponding optimal price values pre-determined for each search term recorded in the search log. The optimal price value corresponding to a search term is determined by looking up the prices of goods that meet specified conditions and were found when searching based on the search term in the past, as recorded in the search log. The module provides a method to determine the absolute value of the difference between the price contained in each searched product information and the optimal price found, and to sort and provide the searched product information in ascending order of the determined absolute values. The optimal price corresponding to the pre-determined search term is the price that the user considers reasonable for the products found using that search term. 7. The apparatus as described in claim 6, wherein the optimal value determination module is specifically configured to: search the search log for product information that meets specified conditions in the product information searched each time a search was performed based on the search term in the past, wherein the product information that meets the specified conditions includes product information clicked by the user; and determine the optimal price value corresponding to the search term based on the price contained in the searched product information. 8. The apparatus as described in claim 7, wherein the optimal value determination module is specifically configured to, based on the price contained in the found product information, use a mixture of Gaussian models to fit the probability density function of the price contained in the product information searched based on the search term, and determine the price corresponding to the maximum probability density as the optimal price value corresponding to the search term based on the fitted probability density function. 9. The apparatus as described in claim 8, wherein the optimal value determination module specifically comprises: An initialization unit is configured to determine the total average price of the found product information based on the prices contained in the found product information; determine the total standard deviation of the prices contained in the found product information based on the prices contained in the found product information and the determined total average price; multiply the determined total average price by a first set parameter as a first average price; multiply the determined total average price by a second set parameter as a second average price; wherein the first set parameter is greater than 1 and the second set parameter is less than 1; and use the determined total standard deviation as both the first standard deviation and the second standard deviation. The training unit is used to determine the median value corresponding to each price in each found product information using a formula, where p is the initially set weight, _mu1 is the first average, _mu2 is the second average, _σ1 is the first standard deviation, _σ2 is the second standard deviation, and x... Let _i represent the price contained in the i-th product information found, and gamma(i) represent the median value corresponding to the price contained in the i-th product information found. Based on the determined median value corresponding to each price, a first average value is re-determined using a formula, and a second average value is re-determined using a formula. Based on the re-determined first average value and second average value, a first standard deviation is re-determined using a formula, and a second standard deviation is re-determined using a formula. The weights are re-determined using a formula, where n represents the number of product information found. It is determined whether the re-determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions. If the determination result is negative, based on the re-determined first average value, second average value, first standard deviation, second standard deviation, and weights, the median value corresponding to each price is re-determined, and based on each re-determined median value, the first average value, second average value, first standard deviation, second standard deviation, and weights are further determined until the determined first average value, second average value, first standard deviation, and second standard deviation meet the set conditions. The function determination unit is used to determine a function as the probability density function of the price contained in the product information searched based on the search term, when the training unit determines that the re-determined first average, second average, first standard deviation, and second standard deviation meet the set conditions. 10. The apparatus of claim 9, wherein the training unit is specifically configured to: determine that the newly determined first average, second average, first standard deviation, and second standard deviation satisfy the set conditions when the difference between the newly determined first average and the previously determined first average is within a set range, or when the difference between the newly determined second average and the previously determined second average is within a set range, or when the difference between the newly determined first standard deviation and the previously determined first standard deviation is within a set range, or when the difference between the newly determined second standard deviation and the previously determined second standard deviation is within a set range.