Vehicle quality inspection method and device using text mining

Abstract

The present invention relates to a vehicle quality inspection method using text mining, comprising: a step in which an electronic device performs preprocessing on input data composed of text; a step in which the electronic device performs text mining on the pre-processed input data; a step in which the electronic device marks one or more mimicry words or one or more morphemes extracted from the preprocessed input data as text mining is performed; and a step in which the electronic device checks the quality of the vehicle on the basis of the marker. The method for checking the quality of the vehicle using text mining of the present invention can also be applied in other embodiments.

Description

Technical Field

[0001] The invention relates to a vehicle quality inspection method and device using text mining. Background Art

[0002] The domestic automobile industry in South Korea is trying to prevent vehicle quality problems in advance based on the analysis of complaint data. In particular, it is trying to prevent recalls and stability-related customer campaigns by discovering stability problems in advance, and focusing on minimizing the cost losses caused by recalls and campaigns. To this end, a complaint data management system led by original equipment manufacturers (OEMs) is being established for the entire vehicle.

[0003] However, currently, since detailed information of quality problems is stored in a non-standardized text form, there is a problem that it is not easy to check the dissatisfaction of core customers. For example, currently detailed information related to quality problems is manually made in files such as Excel, so there is the inconvenience that engineers need to check files such as Excel one by one to check customer dissatisfaction. Summary of the invention

[0004] Problem that the invention aims to solve

[0005] In order to solve this existing problem, an embodiment of the present invention provides a vehicle quality inspection method and device using text mining, which can analyze complaint data and inspect the quality of vehicles through multiple algorithms for text mining, and can update reference data by labeling the analysis results.

[0006] Means used to solve problems

[0007] According to an embodiment of the present invention, the vehicle quality inspection method using text mining is characterized in that it includes: a step in which an electronic device performs preprocessing on input data consisting of text; a step in which the electronic device performs text mining on the preprocessed input data; a step in which the electronic device marks one or more onomatopoeia or one or more morphemes extracted from the preprocessed input data as the text mining is performed; and a step in which the electronic device inspects the vehicle quality based on the marking.

[0008] In addition, the vehicle quality inspection method using text mining is characterized in that, after the step of performing text mining, it also includes: the electronic device checks the quality of the vehicle based on basic errors detected from the preprocessed input data as the text mining is performed.

[0009] In addition, the vehicle quality inspection method using text mining is characterized in that the step of performing text mining includes the step of extracting the one or more onomatopoeias from the pre-processed input data.

[0010] In addition, the vehicle quality inspection method using text mining is characterized in that the step of performing text mining includes the step of extracting the one or more morphemes from the pre-processed input data.

[0011] In addition, the vehicle quality inspection method using text mining is characterized in that the step of marking includes: the step of checking one or more similar keywords related to the one or more extracted onomatopoeias from pre-stored reference data; and the step of marking using representative keywords corresponding to the similar keywords.

[0012] In addition, the vehicle quality inspection method using text mining is characterized in that the step of marking includes: the step of checking one or more similar keywords related to the one or more morphemes included in the morpheme set from pre-stored reference data; and the step of marking using representative keywords corresponding to the similar keywords.

[0013] In addition, the vehicle quality inspection method using text mining is characterized in that, after the step of marking, it also includes a step in which the electronic device uses the one or more onomatopoeias or the one or more morphemes that end the marking to update the reference data.

[0014] In addition, the vehicle quality inspection method using text mining is characterized in that the step of inspecting the quality of the vehicle based on basic errors is a step of inspecting the quality of the vehicle based on the warning lights of the vehicle.

[0015] In addition, the vehicle quality inspection method using text mining is characterized in that the step of inspecting the vehicle quality based on marking is a step of inspecting the quality of the vehicle based on the representative keywords marked in the one or more onomatopoeias related to the noise generated by the vehicle.

[0016] In addition, the vehicle quality inspection method using text mining is characterized in that the step of inspecting the quality of the vehicle based on tags is a step of inspecting the quality of the vehicle related to the operating mode, performance and other errors of the vehicle including oil leakage based on the representative keywords marked on the one or more morphemes.

[0017] In addition, the vehicle quality inspection device using text mining according to an embodiment of the present invention is characterized in that it includes: a control unit, which performs preprocessing on input data consisting of text, and marks one or more onomatopoeias or one or more morphemes extracted from the input data by performing text mining on the preprocessed input data, and inspects the quality of the vehicle based on the marking; and a display unit, which is used to display information related to the quality of the inspected vehicle.

[0018] Furthermore, the vehicle quality inspection device using text mining is characterized in that a control unit inspects the quality of the vehicle based on basic errors inspected from the preprocessed input data as the text mining is performed.

[0019] Furthermore, the vehicle quality inspection device using text mining is characterized in that the control unit extracts the one or more onomatopoeias from the pre-processed input data.

[0020] Furthermore, the vehicle quality inspection device using text mining is characterized in that the control unit extracts the one or more morphemes from the pre-processed input data.

[0021] In addition, the vehicle quality inspection device using text mining is characterized in that the control unit checks one or more similar keywords related to the extracted one or more onomatopoeias from pre-stored reference data, and marks them using representative keywords corresponding to the similar keywords.

[0022] In addition, the vehicle quality inspection device using text mining is characterized in that the control unit generates a set of the one or more morphemes extracted, and checks one or more similar keywords related to the one or more morphemes from the reference data, and uses representative keywords corresponding to the similar keywords for marking.

[0023] Furthermore, the vehicle quality inspection device using text mining is characterized in that the control unit updates the reference data using the one or more onomatopoeias or the one or more morphemes that end the tag.

[0024] Furthermore, the vehicle quality inspection device using text mining is characterized in that the control unit inspects the quality of the vehicle based on a warning light of the vehicle.

[0025] Furthermore, the vehicle quality inspection device using text mining is characterized in that the control unit inspects the quality of the vehicle based on the representative keywords tagged to the one or more onomatopoeias related to the noise generated by the vehicle.

[0026] In addition, the vehicle quality inspection device using text mining is characterized in that the control unit inspects the quality of the vehicle related to the vehicle's operating mode, performance, and other errors including oil leakage based on the representative keywords marked on the one or more morphemes.

[0027] Effects of the Invention

[0028] As described above, according to the vehicle quality inspection method and device using text mining of the present invention, complaint data are analyzed by multiple algorithms for text mining to check the quality of the vehicle, and reference data are updated by marking the analysis results, thereby eliminating the inconvenience of engineers having to check files such as Excel one by one to check customer dissatisfaction, and gradually improving the accuracy of the analysis results of the complaint data by updating the reference data. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a diagram illustrating a main configuration of an electronic device for inspecting vehicle quality using text mining according to an embodiment of the present invention.

[0030] Figure 2 is a flowchart for explaining a method for checking vehicle quality by using text mining according to an embodiment of the present invention.

[0031] Figure 3 is a detailed flowchart for explaining in more detail the method for checking vehicle quality by using text mining according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Figure 1 The detailed description disclosed below is intended to describe exemplary embodiments of the present invention, but is not intended to represent the only implementation method that can implement the present invention. In order to clearly illustrate the present invention, parts that are not related to the description may be omitted from the drawings, and the same reference numerals may be used for the same or similar elements throughout the specification.

[0033] Figure 1 is a diagram illustrating a main configuration of an electronic device for inspecting vehicle quality using text mining according to an embodiment of the present invention.

[0034] Reference Figure 1, the electronic device 100 according to the present invention may include a communication unit 110 , an input unit 120 , a display unit 130 , a memory 140 and a control unit 150 .

[0035] The communication unit 110 may receive input data from an external device through communication with the external device (not shown). At this time, the input data may refer to complaint data in text form. To this end, the communication unit 110 may perform wireless communication such as LTE (long term evolution), Wi-Fi (wireless fidelity), etc. with the external device.

[0036] The input unit 120 generates input data in response to user input of the electronic device 100. To this end, the input unit 120 may include more than one input device. The input unit 120 may include a keyboard, a keypad, a dome switch, a touch panel, a touch key, a mouse, and the like.

[0037] The display unit 130 is used to display display data based on the operation of the electronic device 100. The display unit 130 includes a liquid crystal display (LCD), a light emitting diode (LED) display, an organic LED (OLED) display, a micro electro mechanical system (MEMS) display, and an electronic paper display. The display unit 130 can be combined with the input unit 120 to realize a touch screen.

[0038] The memory 140 may store an action program for causing the electronic device 100 to perform an action, and may store a plurality of algorithms that serve as a basis for checking the quality of the vehicle by analyzing the input data. More specifically, the memory 140 may store various algorithms for analyzing the input data, including the Chat GPT algorithm (hereinafter referred to as the first algorithm) and the Word2Vec algorithm (hereinafter referred to as the second algorithm). In addition, the memory 140 may store reference data for mapping one or more similar keywords, vehicle quality, etc. to representative keywords and lower levels of representative keywords in order to check the quality of the vehicle.

[0039] The control unit 150 may receive complaint data (hereinafter referred to as input data) implemented in a text form from an engineer through the input unit 120 in order to check the quality of the vehicle. In addition, the control unit 150 may receive input data from an external device (not shown) through the communication unit 110. The control unit 150 performs data preprocessing for the input data. The control unit 150 may perform data preprocessing by removing spaces, special characters, etc. from the input data.

[0040] The control unit 150 performs text mining on the input data that has finished preprocessing. More specifically, the control unit 150 calls multiple algorithms stored in the memory 140 to analyze the input data that has finished preprocessing. The multiple algorithms can be Chat GPT algorithm (hereinafter referred to as the first algorithm) and Word2Vec algorithm (hereinafter referred to as the second algorithm). The control unit 150 performs analysis on the input data according to five categories, and the five categories include noise generated by the vehicle, the operating mode of the vehicle, the performance of the vehicle, basic errors such as vehicle warning lights (default error) and other errors, such as wear, oil leakage, wear, small bearings, vibration, cold, heat, upgrade, poor fastening, friction, positioning, side slip, excessive clearance, steel plate noise (steel plate noise) and twisted handle (twisted handle).

[0041] The control unit 150 uses the first algorithm to extract one or more onomatopeia included in the preprocessed input data so as to check (confirm) the noise in the five types. The control unit 150 marks each onomatopeia. If the control unit 150 uses the first algorithm to extract onomatopeia such as "beep", "beep beep", "buzz buzz buzz" from the preprocessed input data, one or more similar keywords related to each onomatopeia extracted from the pre-stored reference data can be checked. That is, similar keywords related to the onomatopeia "beep" can be confirmed as "beep beep", "beep beep", "bubble", "chi chi"; similar keywords related to the onomatopeia "beep beep" can be confirmed as "du lu lu lu", "gur lu lu lu"; similar keywords related to the onomatopeia "buzz buzz buzz" can be confirmed as "buzz buzz buzz", "oh buzz", "buzz bu bu bu", "buzz bu bu".

[0042] In addition, the control unit 150 can separate the extracted onomatopeia in units of characters (letters, words), and apply an edit distance algorithm to confirm similar keywords as representative keywords, wherein the similar keywords are similar keywords that are closest to the onomatopeia separated in units of characters in the predefined similar keywords. For example, the representative keywords related to the onomatopeia "beep" can be used to check "beep", the representative keywords related to the onomatopeia "dudududu" can be used to check "duludulu", and the representative keywords related to the onomatopeia "buzz buzz buzz buzz" can be used to check "buzz buzz buzz buzz".

[0043] The control unit 150 identifies a representative keyword corresponding to one or more similar keywords that have been checked, thereby marking the extracted onomatopeia. In addition, the control unit 150 can update the reference data using the representative keyword marked on each onomatopeia. Thus, the control unit 150 can match one or more onomatopeia that have been checked to the representative keyword, thereby directly extracting the representative keyword when the same onomatopeia is checked in the future.

[0044] In addition, the control unit 150 generates a set of morphemes by extracting one or more morphemes included in the preprocessed input data to check the operation mode of the vehicle in five types, the performance of the vehicle, and other errors. The control unit 150 marks each morpheme included in the set of morphemes. The control unit 150 assigns a score based on the frequency of occurrence of multiple characters constituting a word by recording a cohesion score for the preprocessed input data, and can extract a word only when the minimum frequency of the word is 5 or more and the bonding frequency between two syllables when reading the word from left to right is 5 or more.

[0045] The control unit 150 checks one or more similar keywords related to one or more morphemes, and marks them by confirming representative keywords corresponding to the checked one or more similar keywords. For example, one or more similar keywords can be checked from the pre-stored reference data, and the one or more similar keywords are related to the morphemes "operating", "dragged", and "stiffened" extracted from the morpheme set generated in the preprocessed input data. That is, similar keywords related to the morpheme "operating" may be "operating" and "turning"; similar keywords related to the morpheme "dragged" may be "drag and unsmooth", "drag"; similar keywords related to the morpheme "stiffened" may be "stiff" and "stiff".

[0046] In addition, the control unit 150 may apply the second algorithm to check representative keywords among similar keywords. For example, the representative keywords related to the morpheme "operating" may be used to confirm "operating", the representative keywords related to the morpheme "drag" may be used to confirm "stuck", and the representative keywords related to the morpheme "become stiff" may be used to confirm "stiff".

[0047] Therefore, the control unit 150 marks each morpheme with the checked representative keyword. In addition, the control unit 150 can update the reference data with the representative keyword marked in each morpheme. Thus, the control unit 150 can match the checked representative keywords of one or more morphemes to the representative keyword, thereby directly extracting the representative keyword when the same morpheme is checked in the future.

[0048] If the control unit 150 detects a basic error in the preprocessed input data, it can be marked. For example, if the control unit 150 confirms that a warning light code has been determined in the preprocessed input data, or there are keywords such as warning light on, warning words and sentences on, etc. in the article, the keyword can be extracted and marked. The control unit 150 can mark C2401 in the preprocessed input data and can detect the basic error of the vehicle corresponding to the warning light code.

[0049] When one or more of onomatopeia, morpheme and base error are detected from the pre-processed input data, the control unit 150 can detect the part with quality problem of the vehicle by confirming the mark. The control unit 150 displays the quality inspection result of the vehicle on the display unit 130 .

[0050] Figure 2 is a flowchart for explaining a method for checking vehicle quality by using text mining according to an embodiment of the present invention.

[0051] Reference Figure 2 In step 201, the control unit 150 may receive complaint data (hereinafter referred to as input data) implemented in text form from an engineer through the input unit 120 in order to check the quality of the vehicle. In addition, the control unit 150 may receive input data from an external device (not shown) through the communication unit 110.

[0052] In step 203, the control unit 150 performs data preprocessing on the input data. The control unit 150 may perform data preprocessing by removing spaces, special characters, etc. from the input data.

[0053] In step 205, the control unit 150 performs text mining on the pre-processed input data. Figure 3 A more detailed explanation is given. Figure 3is a detailed flowchart for explaining in more detail the method for checking vehicle quality by using text mining according to an embodiment of the present invention.

[0054] Reference Figure 3 In step 301, the control unit 150 Figure 2 The input data preprocessed in step 205 is analyzed. To this end, the control unit 150 can perform the analysis of the input data by calling multiple algorithms stored in the memory 140, and the multiple algorithms can be Chat GPT algorithm (hereinafter referred to as the first algorithm) and Word2Vec algorithm (hereinafter referred to as the second algorithm). The control unit 150 performs the analysis on the input data according to five types, and the five types include noise generated by the vehicle, the operating mode of the vehicle, the performance of the vehicle, basic errors such as vehicle warning lights, and other errors, such as wear, oil leakage, wear, small bearings, vibration, cold, heat, upgrade, poor fastening, friction, positioning, side slip, excessive clearance, steel plate noise, and twisted handles.

[0055] In step 303, the control unit 150 uses the first algorithm to check whether one or more onomatopoeia is included in the preprocessed input data, so as to confirm the noise in the five types. If it is confirmed in step 303 that one or more onomatopoeia is included in the preprocessed input data, step 305 is executed. In step 305, the control unit 150 extracts one or more onomatopoeia included in the preprocessed input data, and executes step 307. On the contrary, if it is confirmed that one or more onomatopoeia is not included in the preprocessed input data, the control unit 150 executes step 309.

[0056] In step 307, the control unit 150 performs labeling on each onomatopoeia. More specifically, as shown in Table 1 below, if the control unit 150 uses the first algorithm to extract onomatopoeia such as "beep", "beep beep", "buzz buzz buzz" from the preprocessed input data, one or more similar keywords related to each onomatopoeia extracted from the pre-stored reference data can be checked. That is, similar keywords related to the onomatopoeia "beep" can be confirmed as "beep beep", "beep beep", "buzz beep", "chi chi", similar keywords related to the onomatopoeia "beep beep" can be confirmed as "buzzle beep", "gurgle lulu", and similar keywords related to the onomatopoeia "buzz buzz buzz" can be confirmed as "buzz buzz buzz", "oh buzz", "wool wool buzz", and "wool woo buzz".

[0057] In addition, the control unit 150 can separate the extracted onomatopeia in units of characters, and apply an edit distance algorithm to confirm similar keywords as representative keywords, wherein the similar keywords are similar keywords that are closest to the onomatopeia separated in units of characters in the predefined similar keywords. For example, "beep" can be checked using a representative keyword related to the onomatopeia "beep", "duludulu" can be checked using a representative keyword related to the onomatopeia "dududududu", and "buzzbuzzbuzz" can be checked using a representative keyword related to the onomatopeia "buzzbuzzbuzz".

[0058] The control unit 150 identifies a representative keyword corresponding to one or more similar keywords that have been checked, thereby labeling the extracted onomatopeia. That is, the control unit 150 can update the reference data using the representative keyword checked when the representative keyword and the label are the same. Thus, the control unit 150 can match the one or more onomatopeia that have been checked to the representative keyword, thereby directly extracting the representative keyword when the same onomatopeia is checked in the future.

[0059] Table 1

[0060] GPT Extraction Similar keywords in reference data Results of applying edit distance Mark Beep Beep, beep, beep, chirp Beep Beep Beep beep beep Gulu gulu Duludulu Duludulu Buzz buzz buzz Buzz, oh, hum, hum, hum Buzz buzz buzz Buzz buzz buzz

[0061] In step 309, the control unit 150 checks whether one or more morphemes are included in the preprocessed input data, so as to check the operation mode of the vehicle, the performance of the vehicle and other errors in the five types. In step 309, if it is checked that one or more morphemes are included in the preprocessed input data, the control unit 150 forms a set of morphemes by extracting one or more morphemes included in the preprocessed input data in step 311, and executes step 313. For example, "operating" can be checked using a representative keyword related to the morpheme "operating", "stuck" can be checked using a representative keyword related to the morpheme "dragged", and "stiff" can be checked using a representative keyword related to the morpheme "becoming stiff". On the contrary, if it is not in a state where one or more morphemes are included in the preprocessed input data, the control unit 150 executes step 315.

[0062] In step 313, the control unit 150 marks each morpheme. More specifically, the set of morphemes generated based on the preprocessed input data is shown in Table 2 below. At this time, the control unit 150 records the cohesion score for the preprocessed input data, assigning scores based on the frequency of occurrence of multiple characters constituting the word, and extracting the word only when the minimum frequency of the word is 5 or more and the bonding frequency between two syllables when reading the word from left to right is 5 or more.

[0063] Table 2

[0064]

[0065] The control unit 150 checks one or more similar keywords related to one or more morphemes, and labels by checking representative keywords corresponding to the checked one or more similar keywords. That is, the control unit 150 can update the reference data using the representative keyword checked when the representative keyword is the same as the label.

[0066] More specifically, as shown in Table 3 below, the control unit 150 labels the extracted onomatopeia by confirming a representative keyword corresponding to one or more similar keywords checked from the preprocessed input data. That is, the control unit 150 can update the reference data using the representative keyword checked when the representative keyword is the same as the label. Thus, the control unit 150 can match the checked one or more onomatopeia to the representative keyword, thereby directly extracting the representative keyword when the same onomatopeia is checked in the future.

[0067] For example, one or more similar keywords can be checked from the pre-stored reference data, and the one or more similar keywords are similar keywords related to the morphemes "operating", "delayed", and "stiffened" extracted from the set of morphemes. That is, similar keywords related to the morpheme "operating" can be "operating", "turning", similar keywords related to the morpheme "delayed" can be "delayed and unsmooth", "delayed", and similar keywords related to the morpheme "stiffened" can be "stiff", "stiff".

[0068] In addition, the control unit 150 may apply the second algorithm to check representative keywords among similar keywords. For example, the representative keywords related to the morpheme "operating" may be used to check "operating", the representative keywords related to the morpheme "drag" may be used to check "stuck", and the representative keywords related to the morpheme "become stiff" may be used to check "stiff".

[0069] Therefore, the control unit 150 marks the checked representative keywords on each morpheme. In addition, the control unit 150 can update the reference data using the representative keywords marked on each morpheme. Thus, the control unit 150 can match one or more checked morphemes to the representative keywords, so that the representative keywords can be directly extracted when the same morphemes are checked in the future.

[0070] Table 3

[0071]

[0072] Next, in step 315, the control unit 150 analyzes the preprocessed input data to check for basic errors in five types. If a basic error is detected in the preprocessed input data, the control unit 150 may mark it in step 317. For example, as shown in Table 4 below, if a confirmed warning light code or the like is checked in the preprocessed input data, or there are keywords such as warning light on, warning words on, etc. in the article, the control unit 150 may extract these and mark them. That is, as shown in Table 4, if a warning light code such as C2401 is checked in the preprocessed input data, the control unit 150 may mark C2401 in the preprocessed input data, and may check for a basic error of the vehicle corresponding to the warning light code.

[0073] Table 4

[0074]

[0075] On the contrary, if no base error is detected in the preprocessed input data, the control unit 150 may check the quality of the vehicle by executing step 319. That is, the control unit 150 may check the quality of the vehicle based on the marking performed in step 307, step 313, or step 317. That is, if no onomatopoeia, morpheme, or base error is detected in the preprocessed input data, the control unit 150 may confirm that there is no quality problem in the vehicle, and if one or more of the onomatopoeia, morpheme, and base error is detected, the part with a problem in the quality of the vehicle may be checked by checking the marking.

[0076] Next, in step 207, the control unit 150 will Figure 3 The quality inspection result inspected in step 319 is displayed on the display unit 130.

[0077] The embodiments of the present invention shown in this specification and the drawings are only specific examples proposed to easily illustrate the technical content of the present invention and to facilitate the understanding of the present invention, and are not intended to limit the scope of the present invention. Therefore, the scope of the present invention should be interpreted as including all changes or modifications based on the technical ideas of the present invention in addition to the embodiments disclosed in this specification.

Claims

1. A vehicle quality inspection method using text mining, characterized in that: include: The electronic device performs a pre-processing step on the input data consisting of text; The electronic device performs text mining on the pre-processed input data; The electronic device tags one or more onomatopoeias or one or more morphemes extracted from the pre-processed input data as the text mining is performed; as well as The step of the electronic device checking the quality of the vehicle based on the marking.

2. The vehicle quality inspection method using text mining according to claim 1, characterized in that: After performing the text mining steps, it also includes: The electronic device checks the quality of the vehicle based on underlying errors checked from the pre-processed input data as the text mining is performed.

3. The vehicle quality inspection method using text mining according to claim 2 is characterized in that: The steps to perform text mining include: The step of extracting the one or more onomatopeia from the pre-processed input data.

4. The vehicle quality inspection method using text mining according to claim 3 is characterized in that: The steps to perform text mining include: The step of extracting the one or more morphemes from the pre-processed input data.

5. The vehicle quality inspection method using text mining according to claim 4 is characterized in that: The steps for labeling include: A step of checking pre-stored reference data for one or more similar keywords associated with the extracted one or more onomatopoeias; and The step of tagging is performed using representative keywords corresponding to the similar keywords.

6. The vehicle quality inspection method using text mining according to claim 5, characterized in that: The steps for labeling include: A step of generating a set of the one or more extracted morphemes; A step of checking the reference data for one or more similar keywords associated with the one or more morphemes included in the set of morphemes; and The step of tagging is performed using representative keywords corresponding to the similar keywords.

7. The vehicle quality inspection method using text mining according to claim 6, characterized in that: After the marking step, it also includes: The electronic device updates the reference data using the one or more onomatopeia or the one or more morphemes that have been marked.

8. The vehicle quality inspection method using text mining according to claim 2, characterized in that: The step of checking the quality of the vehicle based on the basic error is a step of checking the quality of the vehicle based on a warning light of the vehicle.

9. The vehicle quality inspection method using text mining according to claim 7, characterized in that: The step of checking the quality of the vehicle based on the mark is a step of checking the quality of the vehicle based on the representative keywords marked on the one or more onomatopoeias related to the noise generated by the vehicle.

10. The vehicle quality inspection method using text mining according to claim 7, characterized in that: The step of checking the quality of the vehicle based on the mark is a step of checking the quality of the vehicle related to the operation mode, performance and other errors including oil leakage of the vehicle based on the representative keywords marked on the one or more morphemes.

11. A vehicle quality inspection device using text mining, characterized in that: include: a control unit that performs preprocessing on input data consisting of text, and marks one or more onomatopoeias or one or more morphemes extracted from the input data by performing text mining on the preprocessed input data, and checks the quality of the vehicle based on the marks; as well as The display unit displays information related to the inspected quality of the vehicle.

12. The vehicle quality inspection device using text mining according to claim 11, characterized in that: The control portion checks the quality of the vehicle based on basic errors checked from the preprocessed input data as the text mining is performed.

13. The vehicle quality inspection device using text mining according to claim 12, characterized in that: The control unit extracts the one or more onomatopeia from the pre-processed input data.

14. The vehicle quality inspection device using text mining according to claim 13, characterized in that: The control unit extracts the one or more morphemes from the preprocessed input data.

15. The vehicle quality inspection device using text mining according to claim 14, characterized in that: The control section checks one or more similar keywords related to the extracted one or more onomatopoeias from pre-stored reference data, and tags them with representative keywords corresponding to the similar keywords.

16. The vehicle quality inspection device using text mining according to claim 15, characterized in that: The control unit generates a set of the extracted one or more morphemes, checks the reference data for one or more similar keywords related to the one or more morphemes, and tags them with representative keywords corresponding to the similar keywords.

17. The vehicle quality inspection device using text mining according to claim 16, characterized in that: The control unit updates the reference data using the one or more onomatopeia or the one or more morphemes that have been marked.

18. The vehicle quality inspection device using text mining according to claim 12, characterized in that: The control portion checks the quality of the vehicle based on a warning light of the vehicle.

19. The vehicle quality inspection device using text mining according to claim 17, characterized in that: The control portion checks the quality of the vehicle based on the representative keyword tagged to the one or more onomatopoeia tags related to the noise generated by the vehicle.

20. The vehicle quality inspection device using text mining according to claim 17, characterized in that: The control portion checks the quality of the vehicle related to the operation mode, performance, and other errors including oil leakage of the vehicle based on the representative keywords marked on the one or more morphemes.

Images